redis iam check

feat: support Chroma vector store (#5015 )
Fix: google storage init with sa and download (#5054 )
2024-06-14 03:07:17 +08:00 · 2024-06-13 18:02:18 +08:00 · 2024-06-13 17:36:34 +08:00 · 2024-06-13 17:36:14 +08:00 · 2024-06-13 17:36:01 +08:00 · 2024-06-13 17:16:59 +08:00
194 changed files with 12447 additions and 762 deletions
@@ -14,7 +14,6 @@ concurrency:

 jobs:
  test:
-    name: API Tests
    runs-on: ubuntu-latest
    strategy:
      matrix:
@@ -59,7 +58,7 @@ jobs:
      - name: Run Workflow
        run: dev/pytest/pytest_workflow.sh

-      - name: Set up Vector Stores (Weaviate, Qdrant, PGVector, Milvus, PgVecto-RS)
+      - name: Set up Vector Stores (Weaviate, Qdrant, PGVector, Milvus, PgVecto-RS, Chroma)
        uses: hoverkraft-tech/[email protected]
        with:
          compose-file: |
@@ -68,6 +67,7 @@ jobs:
            docker/docker-compose.milvus.yaml
            docker/docker-compose.pgvecto-rs.yaml
            docker/docker-compose.pgvector.yaml
+            docker/docker-compose.chroma.yaml
          services: |
            weaviate
            qdrant
@@ -76,6 +76,82 @@ jobs:
            milvus-standalone
            pgvecto-rs
            pgvector
+            chroma

      - name: Test Vector Stores
        run: dev/pytest/pytest_vdb.sh
+
+  test-in-poetry:
+    name: API Tests
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version:
+          - "3.10"
+          - "3.11"
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Install Poetry
+        uses: abatilo/actions-poetry@v3
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+          cache: 'poetry'
+          cache-dependency-path: |
+            api/pyproject.toml
+            api/poetry.lock
+
+      - name: Poetry check
+        run: poetry check -C api
+
+      - name: Install dependencies
+        run: poetry install -C api --with dev
+
+      - name: Run Unit tests
+        run: poetry run -C api bash dev/pytest/pytest_unit_tests.sh
+
+      - name: Run ModelRuntime
+        run: poetry run -C api bash dev/pytest/pytest_model_runtime.sh
+
+      - name: Run Tool
+        run: poetry run -C api bash dev/pytest/pytest_tools.sh
+
+      - name: Set up Sandbox
+        uses: hoverkraft-tech/[email protected]
+        with:
+          compose-file: |
+            docker/docker-compose.middleware.yaml
+          services: |
+            sandbox
+            ssrf_proxy
+
+      - name: Run Workflow
+        run: poetry run -C api bash dev/pytest/pytest_workflow.sh
+
+      - name: Set up Vector Stores (Weaviate, Qdrant, PGVector, Milvus, PgVecto-RS, Chroma)
+        uses: hoverkraft-tech/[email protected]
+        with:
+          compose-file: |
+            docker/docker-compose.middleware.yaml
+            docker/docker-compose.qdrant.yaml
+            docker/docker-compose.milvus.yaml
+            docker/docker-compose.pgvecto-rs.yaml
+            docker/docker-compose.pgvector.yaml
+            docker/docker-compose.chroma.yaml
+          services: |
+            weaviate
+            qdrant
+            etcd
+            minio
+            milvus-standalone
+            pgvecto-rs
+            pgvector
+            chroma
+
+      - name: Test Vector Stores
+        run: poetry run -C api bash dev/pytest/pytest_vdb.sh
@@ -23,16 +23,20 @@ jobs:
      - name: Checkout code
        uses: actions/checkout@v4

+      - name: Install Poetry
+        uses: abatilo/actions-poetry@v3
+
      - name: Set up Python ${{ matrix.python-version }}
        uses: actions/setup-python@v5
        with:
          python-version: ${{ matrix.python-version }}
-          cache: 'pip'
+          cache: 'poetry'
          cache-dependency-path: |
-            ./api/requirements.txt
+            api/pyproject.toml
+            api/poetry.lock

      - name: Install dependencies
-        run: pip install -r ./api/requirements.txt
+        run: poetry install -C api

      - name: Set up Middleware
        uses: hoverkraft-tech/[email protected]
@@ -50,4 +54,4 @@ jobs:
      - name: Run DB Migration
        run: |
          cd api
-          flask db upgrade
+          poetry run python -m flask db upgrade
@@ -24,6 +24,9 @@ jobs:
        with:
          files: api/**

+      - name: Install Poetry
+        uses: abatilo/actions-poetry@v3
+
      - name: Set up Python
        uses: actions/setup-python@v5
        if: steps.changed-files.outputs.any_changed == 'true'
@@ -32,15 +35,15 @@ jobs:

      - name: Python dependencies
        if: steps.changed-files.outputs.any_changed == 'true'
-        run: pip install ruff dotenv-linter
+        run: poetry install -C api --only lint

      - name: Ruff check
        if: steps.changed-files.outputs.any_changed == 'true'
-        run: ruff check ./api
+        run: poetry run -C api ruff check --preview ./api

      - name: Dotenv check
        if: steps.changed-files.outputs.any_changed == 'true'
-        run: dotenv-linter ./api/.env.example ./web/.env.example
+        run: poetry run -C api dotenv-linter ./api/.env.example ./web/.env.example

      - name: Lint hints
        if: failure()
@@ -149,6 +149,7 @@ docker/volumes/qdrant/*
 docker/volumes/etcd/*
 docker/volumes/minio/*
 docker/volumes/milvus/*
+docker/volumes/chroma/*

 sdks/python-client/build
 sdks/python-client/dist
@@ -36,6 +36,7 @@
  <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>
+  <a href="./README_AR.md"><img alt="README بالعربية" src="https://img.shields.io/badge/العربية-d9d9d9"></a>
 </p>


@@ -0,0 +1,225 @@
+![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 Cloud</a> ·
+  <a href="https://docs.dify.ai/getting-started/install-self-hosted">الاستضافة الذاتية</a> ·
+  <a href="https://docs.dify.ai">التوثيق</a> ·
+  <a href="https://cal.com/guchenhe/60-min-meeting">استفسارات الشركات</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="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
+  <a href="./README_AR.md"><img alt="README بالعربية" src="https://img.shields.io/badge/العربية-d9d9d9"></a>
+</p>
+
+<div style="text-align: right;">
+مشروع Dify هو منصة تطوير تطبيقات الذكاء الصناعي مفتوحة المصدر. تجمع واجهته البديهية بين سير العمل الذكي بالذكاء الاصطناعي وخط أنابيب RAG وقدرات الوكيل وإدارة النماذج وميزات الملاحظة وأكثر من ذلك، مما يتيح لك الانتقال بسرعة من المرحلة التجريبية إلى الإنتاج. إليك قائمة بالميزات الأساسية:
+</br> </br>
+
+**1. سير العمل**:  قم ببناء واختبار سير عمل الذكاء الاصطناعي القوي على قماش بصري، مستفيدًا من جميع الميزات التالية وأكثر.
+
+  https://github.com/langgenius/dify/assets/13230914/356df23e-1604-483d-80a6-9517ece318aa
+
+
+**2. الدعم الشامل للنماذج**: تكامل سلس مع مئات من LLMs الخاصة / مفتوحة المصدر من عشرات من موفري التحليل والحلول المستضافة ذاتيًا، مما يغطي GPT و Mistral و Llama3 وأي نماذج متوافقة مع واجهة OpenAI API. يمكن العثور على قائمة كاملة بمزودي النموذج المدعومين [هنا](https://docs.dify.ai/getting-started/readme/model-providers).
+
+![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)
+
+**3. بيئة التطوير للأوامر**: واجهة بيئة التطوير المبتكرة لصياغة الأمر ومقارنة أداء النموذج، وإضافة ميزات إضافية مثل تحويل النص إلى كلام إلى تطبيق قائم على الدردشة.
+
+**4. خط أنابيب RAG**: قدرات RAG الواسعة التي تغطي كل شيء من استيعاب الوثائق إلى الاسترجاع، مع الدعم الفوري لاستخراج النص من ملفات PDF و PPT وتنسيقات الوثائق الشائعة الأخرى.
+
+**5. قدرات الوكيل**: يمكنك تعريف الوكلاء بناءً على أمر وظيفة LLM أو ReAct، وإضافة أدوات مدمجة أو مخصصة للوكيل. توفر Dify أكثر من 50 أداة مدمجة لوكلاء الذكاء الاصطناعي، مثل البحث في Google و DELL·E وStable Diffusion و WolframAlpha.
+
+**6. الـ LLMOps**: راقب وتحلل سجلات التطبيق والأداء على مر الزمن. يمكنك تحسين الأوامر والبيانات والنماذج باستمرار استنادًا إلى البيانات الإنتاجية والتعليقات.
+
+**7.الواجهة الخلفية (Backend) كخدمة**: تأتي جميع عروض Dify مع APIs مطابقة، حتى يمكنك دمج Dify بسهولة في منطق أعمالك الخاص.
+## مقارنة الميزات
+<table style="width: 100%;">
+  <tr>
+    <th align="center">الميزة</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">نهج البرمجة</td>
+    <td align="center">موجّه لـ تطبيق + واجهة برمجة تطبيق (API)</td>
+    <td align="center">برمجة Python</td>
+    <td align="center">موجه لتطبيق</td>
+    <td align="center">واجهة برمجة تطبيق (API)</td>
+  </tr>
+  <tr>
+    <td align="center">LLMs المدعومة</td>
+    <td align="center">تنوع غني</td>
+    <td align="center">تنوع غني</td>
+    <td align="center">تنوع غني</td>
+    <td align="center">فقط OpenAI</td>
+  </tr>
+  <tr>
+    <td align="center">محرك RAG</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+  </tr>
+  <tr>
+    <td align="center">الوكيل</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">✅</td>
+  </tr>
+  <tr>
+    <td align="center">سير العمل</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+  </tr>
+  <tr>
+    <td align="center">الملاحظة</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">❌</td>
+  </tr>
+  <tr>
+    <td align="center">ميزات الشركات (SSO / مراقبة الوصول)</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+    <td align="center">❌</td>
+    <td align="center">❌</td>
+  </tr>
+  <tr>
+    <td align="center">نشر محلي</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">✅</td>
+    <td align="center">❌</td>
+  </tr>
+</table>
+
+
+## استخدام Dify
+- **سحابة </br>**
+نحن نستضيف [خدمة Dify Cloud](https://dify.ai) لأي شخص لتجربتها بدون أي إعدادات. توفر كل قدرات النسخة التي تمت استضافتها ذاتيًا، وتتضمن 200 أمر GPT-4 مجانًا في خطة الصندوق الرملي.
+
+- **استضافة ذاتية لنسخة المجتمع Dify</br>**
+ابدأ سريعًا في تشغيل Dify في بيئتك باستخدام [دليل البدء السريع](#البدء السريع).
+استخدم [توثيقنا](https://docs.dify.ai) للمزيد من المراجع والتعليمات الأعمق.
+
+- **مشروع Dify للشركات / المؤسسات</br>**
+نحن نوفر ميزات إضافية مركزة على الشركات. [جدول اجتماع معنا](https://cal.com/guchenhe/30min) أو [أرسل لنا بريدًا إلكترونيًا](mailto:[email protected]?subject=[GitHub]Business%20License%20Inquiry) لمناقشة احتياجات الشركات. </br>
+> بالنسبة للشركات الناشئة والشركات الصغيرة التي تستخدم خدمات AWS، تحقق من [Dify Premium على AWS Marketplace](https://aws.amazon.com/marketplace/pp/prodview-t22mebxzwjhu6) ونشرها في شبكتك الخاصة على AWS VPC بنقرة واحدة. إنها عرض AMI بأسعار معقولة مع خيار إنشاء تطبيقات بشعار وعلامة تجارية مخصصة.
+## البقاء قدمًا
+
+قم بإضافة نجمة إلى Dify على GitHub وتلق تنبيهًا فوريًا بالإصدارات الجديدة.
+
+![نجمنا](https://github.com/langgenius/dify/assets/13230914/b823edc1-6388-4e25-ad45-2f6b187adbb4)
+## البداية السريعة
+> قبل تثبيت Dify، تأكد من أن جهازك يلبي الحد الأدنى من متطلبات النظام التالية:
+> 
+>- معالج >= 2 نواة
+>- ذاكرة وصول عشوائي (RAM) >= 4 جيجابايت
+
+</br>
+
+أسهل طريقة لبدء تشغيل خادم Dify هي تشغيل ملف [docker-compose.yml](docker/docker-compose.yaml) الخاص بنا. قبل تشغيل أمر التثبيت، تأكد من تثبيت [Docker](https://docs.docker.com/get-docker/) و [Docker Compose](https://docs.docker.com/compose/install/) على جهازك:
+
+```bash
+cd docker
+docker compose up -d
+```
+بعد التشغيل، يمكنك الوصول إلى لوحة تحكم Dify في متصفحك على [http://localhost/install](http://localhost/install) وبدء عملية التهيئة.
+
+> إذا كنت ترغب في المساهمة في Dify أو القيام بتطوير إضافي، فانظر إلى [دليلنا للنشر من الشفرة (code) المصدرية](https://docs.dify.ai/getting-started/install-self-hosted/local-source-code)
+
+## الخطوات التالية
+
+إذا كنت بحاجة إلى تخصيص التكوين، يرجى الرجوع إلى التعليقات في ملف [docker-compose.yml](docker/docker-compose.yaml) لدينا وتعيين التكوينات البيئية يدويًا. بعد إجراء التغييرات، يرجى تشغيل `docker-compose up -d` مرة أخرى. يمكنك رؤية قائمة كاملة بالمتغيرات البيئية [هنا](https://docs.dify.ai/getting-started/install-self-hosted/environments).
+
+إذا كنت ترغب في تكوين إعداد متوفر بشكل عالي، فهناك [رسوم بيانية Helm](https://helm.sh/) المساهمة من المجتمع تسمح بنشر Dify على Kubernetes.
+
+- [رسم بياني Helm من قبل @LeoQuote](https://github.com/douban/charts/tree/master/charts/dify)
+- [رسم بياني Helm من قبل @BorisPolonsky](https://github.com/BorisPolonsky/dify-helm)
+
+
+## المساهمة
+
+لأولئك الذين يرغبون في المساهمة، انظر إلى [دليل المساهمة](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md) لدينا. 
+في الوقت نفسه، يرجى النظر في دعم Dify عن طريق مشاركته على وسائل التواصل الاجتماعي وفي الفعاليات والمؤتمرات.
+
+
+> نحن نبحث عن مساهمين لمساعدة في ترجمة Dify إلى لغات أخرى غير اللغة الصينية المندرين أو الإنجليزية. إذا كنت مهتمًا بالمساعدة، يرجى الاطلاع على [README للترجمة](https://github.com/langgenius/dify/blob/main/web/i18n/README.md) لمزيد من المعلومات، واترك لنا تعليقًا في قناة `global-users` على [خادم المجتمع على Discord](https://discord.gg/8Tpq4AcN9c).
+
+**المساهمون**
+
+<a href="https://github.com/langgenius/dify/graphs/contributors">
+  <img src="https://contrib.rocks/image?repo=langgenius/dify" />
+</a>
+
+## المجتمع والاتصال
+* [مناقشة Github](https://github.com/langgenius/dify/discussions). الأفضل لـ: مشاركة التعليقات وطرح الأسئلة.
+* [المشكلات على GitHub](https://github.com/langgenius/dify/issues). الأفضل لـ: الأخطاء التي تواجهها في استخدام Dify.AI، واقتراحات الميزات. انظر [دليل المساهمة](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md).
+* [البريد الإلكتروني](mailto:[email protected]?subject=[GitHub]Questions%20About%20Dify). الأفضل لـ: الأسئلة التي تتعلق باستخدام Dify.AI.
+* [Discord](https://discord.gg/FngNHpbcY7). الأفضل لـ: مشاركة تطبيقاتك والترفيه مع المجتمع.
+* [تويتر](https://twitter.com/dify_ai). الأفضل لـ: مشاركة تطبيقاتك والترفيه مع المجتمع.
+
+أو، قم بجدولة اجتماع مباشرة مع أحد أعضاء الفريق:
+
+<table>
+  <tr>
+    <th>نقطة الاتصال</th>
+    <th>الغرض</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>استفسارات الأعمال واقتراحات حول المنتج</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>المساهمات والمشكلات وطلبات الميزات</td>
+  </tr>
+</table>
+
+## تاريخ النجمة
+
+[![Star History Chart](https://api.star-history.com/svg?repos=langgenius/dify&type=Date)](https://star-history.com/#langgenius/dify&Date)
+
+
+## الكشف عن الأمان
+
+لحماية خصوصيتك، يرجى تجنب نشر مشكلات الأمان على GitHub. بدلاً من ذلك، أرسل أسئلتك إلى security@dify.ai وسنقدم لك إجابة أكثر تفصيلاً.
+
+## الرخصة
+
+هذا المستودع متاح تحت [رخصة البرنامج الحر Dify](LICENSE)، والتي تعتبر بشكل أساسي Apache 2.0 مع بعض القيود الإضافية.
@@ -112,6 +112,21 @@ PGVECTOR_USER=postgres
 PGVECTOR_PASSWORD=postgres
 PGVECTOR_DATABASE=postgres

+# Tidb Vector configuration
+TIDB_VECTOR_HOST=xxx.eu-central-1.xxx.aws.tidbcloud.com
+TIDB_VECTOR_PORT=4000
+TIDB_VECTOR_USER=xxx.root
+TIDB_VECTOR_PASSWORD=xxxxxx
+TIDB_VECTOR_DATABASE=dify
+
+# Chroma configuration
+CHROMA_HOST=127.0.0.1
+CHROMA_PORT=8000
+CHROMA_TENANT=default_tenant
+CHROMA_DATABASE=default_database
+CHROMA_AUTH_PROVIDER=chromadb.auth.token_authn.TokenAuthenticationServerProvider
+CHROMA_AUTH_CREDENTIALS=difyai123456
+
 # Upload configuration
 UPLOAD_FILE_SIZE_LIMIT=15
 UPLOAD_FILE_BATCH_LIMIT=5
@@ -17,7 +17,8 @@
                "FLASK_DEBUG": "1",
                "GEVENT_SUPPORT": "True"
            },
-            "console": "integratedTerminal"
+            "console": "integratedTerminal",
+            "python": "${command:python.interpreterPath}"
        },
        {
            "name": "Python: Flask",
@@ -36,7 +37,8 @@
                "--debug"
            ],
            "jinja": true,
-            "justMyCode": true
+            "justMyCode": true,
+            "python": "${command:python.interpreterPath}"
        }
    ]
 }
@@ -17,15 +17,30 @@
   ```bash
   sed -i "/^SECRET_KEY=/c\SECRET_KEY=$(openssl rand -base64 42)" .env
   ```
-4. If you use Anaconda, create a new environment and activate it
+4. Create environment.
+   - Anaconda  
+   If you use Anaconda, create a new environment and activate it
   ```bash
   conda create --name dify python=3.10
   conda activate dify
   ```
+   - Poetry  
+   If you use Poetry, you don't need to manually create the environment. You can execute `poetry shell` to activate the environment.
 5. Install dependencies
+   - Anaconda  
   ```bash
   pip install -r requirements.txt
   ```
+   - Poetry  
+   ```bash
+   poetry install
+   ```
+   In case of contributors missing to update dependencies for `pyproject.toml`, you can perform the following shell instead.
+   ```base
+   poetry shell                                               # activate current environment
+   poetry add $(cat requirements.txt)           # install dependencies of production and update pyproject.toml
+   poetry add $(cat requirements-dev.txt) --group dev    # install dependencies of development and update pyproject.toml
+   ```
 6. Run migrate

   Before the first launch, migrate the database to the latest version.
@@ -1,12 +1,15 @@
 import base64
 import json
 import secrets
+from typing import Optional

 import click
 from flask import current_app
 from werkzeug.exceptions import NotFound

+from constants.languages import languages
 from core.rag.datasource.vdb.vector_factory import Vector
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.models.document import Document
 from extensions.ext_database import db
 from libs.helper import email as email_validate
@@ -17,6 +20,7 @@ from models.dataset import Dataset, DatasetCollectionBinding, DocumentSegment
 from models.dataset import Document as DatasetDocument
 from models.model import Account, App, AppAnnotationSetting, AppMode, Conversation, MessageAnnotation
 from models.provider import Provider, ProviderModel
+from services.account_service import RegisterService, TenantService


@click.command('reset-password', help='Reset the account password.')
@@ -57,7 +61,7 @@ def reset_password(email, new_password, password_confirm):
    account.password = base64_password_hashed
    account.password_salt = base64_salt
    db.session.commit()
-    click.echo(click.style('Congratulations!, password has been reset.', fg='green'))
+    click.echo(click.style('Congratulations! Password has been reset.', fg='green'))


@click.command('reset-email', help='Reset the account email.')
@@ -263,15 +267,15 @@ def migrate_knowledge_vector_database():
                        skipped_count = skipped_count + 1
                        continue
                collection_name = ''
-                if vector_type == "weaviate":
+                if vector_type == VectorType.WEAVIATE:
                    dataset_id = dataset.id
                    collection_name = Dataset.gen_collection_name_by_id(dataset_id)
                    index_struct_dict = {
-                        "type": 'weaviate',
+                        "type": VectorType.WEAVIATE,
                        "vector_store": {"class_prefix": collection_name}
                    }
                    dataset.index_struct = json.dumps(index_struct_dict)
-                elif vector_type == "qdrant":
+                elif vector_type == VectorType.QDRANT:
                    if dataset.collection_binding_id:
                        dataset_collection_binding = db.session.query(DatasetCollectionBinding). \
                            filter(DatasetCollectionBinding.id == dataset.collection_binding_id). \
@@ -284,20 +288,20 @@ def migrate_knowledge_vector_database():
                        dataset_id = dataset.id
                        collection_name = Dataset.gen_collection_name_by_id(dataset_id)
                    index_struct_dict = {
-                        "type": 'qdrant',
+                        "type": VectorType.QDRANT,
                        "vector_store": {"class_prefix": collection_name}
                    }
                    dataset.index_struct = json.dumps(index_struct_dict)

-                elif vector_type == "milvus":
+                elif vector_type == VectorType.MILVUS:
                    dataset_id = dataset.id
                    collection_name = Dataset.gen_collection_name_by_id(dataset_id)
                    index_struct_dict = {
-                        "type": 'milvus',
+                        "type": VectorType.MILVUS,
                        "vector_store": {"class_prefix": collection_name}
                    }
                    dataset.index_struct = json.dumps(index_struct_dict)
-                elif vector_type == "relyt":
+                elif vector_type == VectorType.RELYT:
                    dataset_id = dataset.id
                    collection_name = Dataset.gen_collection_name_by_id(dataset_id)
                    index_struct_dict = {
@@ -305,16 +309,16 @@ def migrate_knowledge_vector_database():
                        "vector_store": {"class_prefix": collection_name}
                    }
                    dataset.index_struct = json.dumps(index_struct_dict)
-                elif vector_type == "pgvector":
+                elif vector_type == VectorType.PGVECTOR:
                    dataset_id = dataset.id
                    collection_name = Dataset.gen_collection_name_by_id(dataset_id)
                    index_struct_dict = {
-                        "type": 'pgvector',
+                        "type": VectorType.PGVECTOR,
                        "vector_store": {"class_prefix": collection_name}
                    }
                    dataset.index_struct = json.dumps(index_struct_dict)
                else:
-                    raise ValueError(f"Vector store {config.get('VECTOR_STORE')} is not supported.")
+                    raise ValueError(f"Vector store {vector_type} is not supported.")

                vector = Vector(dataset)
                click.echo(f"Start to migrate dataset {dataset.id}.")
@@ -501,6 +505,46 @@ def add_qdrant_doc_id_index(field: str):
                    fg='green'))


+@click.command('create-tenant', help='Create account and tenant.')
+@click.option('--email', prompt=True, help='The email address of the tenant account.')
+@click.option('--language', prompt=True, help='Account language, default: en-US.')
+def create_tenant(email: str, language: Optional[str] = None):
+    """
+    Create tenant account
+    """
+    if not email:
+        click.echo(click.style('Sorry, email is required.', fg='red'))
+        return
+
+    # Create account
+    email = email.strip()
+
+    if '@' not in email:
+        click.echo(click.style('Sorry, invalid email address.', fg='red'))
+        return
+
+    account_name = email.split('@')[0]
+
+    if language not in languages:
+        language = 'en-US'
+
+    # generate random password
+    new_password = secrets.token_urlsafe(16)
+
+    # register account
+    account = RegisterService.register(
+        email=email,
+        name=account_name,
+        password=new_password,
+        language=language
+    )
+
+    TenantService.create_owner_tenant_if_not_exist(account)
+
+    click.echo(click.style('Congratulations! Account and tenant created.\n'
+                           'Account: {}\nPassword: {}'.format(email, new_password), fg='green'))
+
+
 def register_commands(app):
    app.cli.add_command(reset_password)
    app.cli.add_command(reset_email)
@@ -508,4 +552,5 @@ def register_commands(app):
    app.cli.add_command(vdb_migrate)
    app.cli.add_command(convert_to_agent_apps)
    app.cli.add_command(add_qdrant_doc_id_index)
+    app.cli.add_command(create_tenant)

@@ -299,6 +299,21 @@ class Config:
        self.PGVECTOR_PASSWORD = get_env('PGVECTOR_PASSWORD')
        self.PGVECTOR_DATABASE = get_env('PGVECTOR_DATABASE')

+        # tidb-vector settings
+        self.TIDB_VECTOR_HOST = get_env('TIDB_VECTOR_HOST')
+        self.TIDB_VECTOR_PORT = get_env('TIDB_VECTOR_PORT')
+        self.TIDB_VECTOR_USER = get_env('TIDB_VECTOR_USER')
+        self.TIDB_VECTOR_PASSWORD = get_env('TIDB_VECTOR_PASSWORD')
+        self.TIDB_VECTOR_DATABASE = get_env('TIDB_VECTOR_DATABASE')
+
+        # chroma settings
+        self.CHROMA_HOST = get_env('CHROMA_HOST')
+        self.CHROMA_PORT = get_env('CHROMA_PORT')
+        self.CHROMA_TENANT = get_env('CHROMA_TENANT')
+        self.CHROMA_DATABASE = get_env('CHROMA_DATABASE')
+        self.CHROMA_AUTH_PROVIDER = get_env('CHROMA_AUTH_PROVIDER')
+        self.CHROMA_AUTH_CREDENTIALS = get_env('CHROMA_AUTH_CREDENTIALS')
+
        # ------------------------
        # Mail Configurations.
        # ------------------------
@@ -15,6 +15,7 @@ from core.errors.error import LLMBadRequestError, ProviderTokenNotInitError
 from core.indexing_runner import IndexingRunner
 from core.model_runtime.entities.model_entities import ModelType
 from core.provider_manager import ProviderManager
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.extractor.entity.extract_setting import ExtractSetting
 from extensions.ext_database import db
 from fields.app_fields import related_app_list
@@ -476,20 +477,22 @@ class DatasetRetrievalSettingApi(Resource):
    @account_initialization_required
    def get(self):
        vector_type = current_app.config['VECTOR_STORE']
-        if vector_type in {"milvus", "relyt", "pgvector", "pgvecto_rs"}:
-            return {
-                'retrieval_method': [
-                    'semantic_search'
-                ]
-            }
-        elif vector_type in {"qdrant", "weaviate"}:
-            return {
-                'retrieval_method': [
-                    'semantic_search', 'full_text_search', 'hybrid_search'
-                ]
-            }
-        else:
-            raise ValueError("Unsupported vector db type.")
+
+        match vector_type:
+            case VectorType.MILVUS | VectorType.RELYT | VectorType.PGVECTOR | VectorType.TIDB_VECTOR | VectorType.CHROMA:
+                return {
+                    'retrieval_method': [
+                        'semantic_search'
+                    ]
+                }
+            case VectorType.QDRANT | VectorType.WEAVIATE:
+                return {
+                    'retrieval_method': [
+                        'semantic_search', 'full_text_search', 'hybrid_search'
+                    ]
+                }
+            case _:
+                raise ValueError(f"Unsupported vector db type {vector_type}.")


 class DatasetRetrievalSettingMockApi(Resource):
@@ -497,20 +500,22 @@ class DatasetRetrievalSettingMockApi(Resource):
    @login_required
    @account_initialization_required
    def get(self, vector_type):
-        if vector_type in {'milvus', 'relyt', 'pgvector'}:
-            return {
-                'retrieval_method': [
-                    'semantic_search'
-                ]
-            }
-        elif vector_type in {'qdrant', 'weaviate'}:
-            return {
-                'retrieval_method': [
-                    'semantic_search', 'full_text_search', 'hybrid_search'
-                ]
-            }
-        else:
-            raise ValueError("Unsupported vector db type.")
+        match vector_type:
+            case VectorType.MILVUS | VectorType.RELYT | VectorType.PGVECTOR | VectorType.TIDB_VECTOR | VectorType.CHROMA:
+                return {
+                    'retrieval_method': [
+                        'semantic_search'
+                    ]
+                }
+            case VectorType.QDRANT | VectorType.WEAVIATE:
+                return {
+                    'retrieval_method': [
+                        'semantic_search', 'full_text_search', 'hybrid_search'
+                    ]
+                }
+            case _:
+                raise ValueError(f"Unsupported vector db type {vector_type}.")
+

 class DatasetErrorDocs(Resource):
    @setup_required
@@ -1,5 +1,6 @@
 from flask import request
 from flask_restful import marshal, reqparse
+from werkzeug.exceptions import NotFound

 import services.dataset_service
 from controllers.service_api import api
@@ -19,10 +20,12 @@ def _validate_name(name):
    return name


-class DatasetApi(DatasetApiResource):
-    """Resource for get datasets."""
+class DatasetListApi(DatasetApiResource):
+    """Resource for datasets."""

    def get(self, tenant_id):
+        """Resource for getting datasets."""
+
        page = request.args.get('page', default=1, type=int)
        limit = request.args.get('limit', default=20, type=int)
        provider = request.args.get('provider', default="vendor")
@@ -65,9 +68,9 @@ class DatasetApi(DatasetApiResource):
        }
        return response, 200

-    """Resource for datasets."""

    def post(self, tenant_id):
+        """Resource for creating datasets."""
        parser = reqparse.RequestParser()
        parser.add_argument('name', nullable=False, required=True,
                            help='type is required. Name must be between 1 to 40 characters.',
@@ -89,6 +92,31 @@ class DatasetApi(DatasetApiResource):

        return marshal(dataset, dataset_detail_fields), 200

+class DatasetApi(DatasetApiResource):
+    """Resource for dataset."""

-api.add_resource(DatasetApi, '/datasets')
+    def delete(self, _, dataset_id):
+        """
+        Deletes a dataset given its ID.

+        Args:
+            dataset_id (UUID): The ID of the dataset to be deleted.
+
+        Returns:
+            dict: A dictionary with a key 'result' and a value 'success' 
+                  if the dataset was successfully deleted. Omitted in HTTP response.
+            int: HTTP status code 204 indicating that the operation was successful.
+
+        Raises:
+            NotFound: If the dataset with the given ID does not exist.
+        """
+
+        dataset_id_str = str(dataset_id)
+
+        if DatasetService.delete_dataset(dataset_id_str, current_user):
+            return {'result': 'success'}, 204
+        else:
+            raise NotFound("Dataset not found.")
+
+api.add_resource(DatasetListApi, '/datasets')
+api.add_resource(DatasetApi, '/datasets/<uuid:dataset_id>')
@@ -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['voice'] if request.form.get('voice') else app_model.app_model_config.text_to_speech_dict.get('voice'),
+                voice=request.form['voice'] if request.form.get('voice') else None,
                streaming=False
            )

@@ -528,4 +528,3 @@ class BaseAgentRunner(AppRunner):
                return UserPromptMessage(content=prompt_message_contents)
        else:
            return UserPromptMessage(content=message.query)
-         
@@ -107,8 +107,8 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
            usage=LLMUsage.empty_usage()
        )

-        self._stream_generate_routes = self._get_stream_generate_routes()
        self._iteration_nested_relations = self._get_iteration_nested_relations(self._workflow.graph_dict)
+        self._stream_generate_routes = self._get_stream_generate_routes()
        self._conversation_name_generate_thread = None

    def process(self) -> Union[ChatbotAppBlockingResponse, Generator[ChatbotAppStreamResponse, None, None]]:
@@ -410,6 +410,18 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                ingoing_edges.append(edge)

        if not ingoing_edges:
+            # check if it's the first node in the iteration
+            target_node = next((node for node in nodes if node.get('id') == target_node_id), None)
+            if not target_node:
+                return []
+            
+            node_iteration_id = target_node.get('data', {}).get('iteration_id')
+            # get iteration start node id
+            for node in nodes:
+                if node.get('id') == node_iteration_id:
+                    if node.get('data', {}).get('start_node_id') == target_node_id:
+                        return [target_node_id]
+                    
            return []

        start_node_ids = []
@@ -514,6 +526,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                self._task_state.answer += route_chunk.text
                yield self._message_to_stream_response(route_chunk.text, self._message.id)
            else:
+                value = None
                route_chunk = cast(VarGenerateRouteChunk, route_chunk)
                value_selector = route_chunk.value_selector
                if not value_selector:
@@ -525,6 +538,20 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                if route_chunk_node_id == 'sys':
                    # system variable
                    value = self._workflow_system_variables.get(SystemVariable.value_of(value_selector[1]))
+                elif route_chunk_node_id in self._iteration_nested_relations:
+                    # it's a iteration variable
+                    if not self._iteration_state or route_chunk_node_id not in self._iteration_state.current_iterations:
+                        continue
+                    iteration_state = self._iteration_state.current_iterations[route_chunk_node_id]
+                    iterator = iteration_state.inputs
+                    if not iterator:
+                        continue
+                    iterator_selector = iterator.get('iterator_selector', [])
+                    if value_selector[1] == 'index':
+                        value = iteration_state.current_index
+                    elif value_selector[1] == 'item':
+                        value = iterator_selector[iteration_state.current_index] if iteration_state.current_index < len(
+                            iterator_selector) else None
                else:
                    # check chunk node id is before current node id or equal to current node id
                    if route_chunk_node_id not in self._task_state.ran_node_execution_infos:
@@ -554,7 +581,7 @@ class AdvancedChatAppGenerateTaskPipeline(BasedGenerateTaskPipeline, WorkflowCyc
                        else:
                            value = value.get(key)

-                if value:
+                if value is not None:
                    text = ''
                    if isinstance(value, str | int | float):
                        text = str(value)
@@ -1,3 +1,5 @@
+from typing import Optional
+
 from core.app.app_config.features.file_upload.manager import FileUploadConfigManager
 from core.file.message_file_parser import MessageFileParser
 from core.model_manager import ModelInstance
@@ -19,7 +21,7 @@ class TokenBufferMemory:
        self.model_instance = model_instance

    def get_history_prompt_messages(self, max_token_limit: int = 2000,
-                                    message_limit: int = 10) -> list[PromptMessage]:
+                                    message_limit: Optional[int] = None) -> list[PromptMessage]:
        """
        Get history prompt messages.
        :param max_token_limit: max token limit
@@ -28,10 +30,15 @@ class TokenBufferMemory:
        app_record = self.conversation.app

        # fetch limited messages, and return reversed
-        messages = db.session.query(Message).filter(
+        query = db.session.query(Message).filter(
            Message.conversation_id == self.conversation.id,
            Message.answer != ''
-        ).order_by(Message.created_at.desc()).limit(message_limit).all()
+        ).order_by(Message.created_at.desc())
+
+        if message_limit and message_limit > 0:
+            messages = query.limit(message_limit).all()
+        else:
+            messages = query.all()

        messages = list(reversed(messages))
        message_file_parser = MessageFileParser(
@@ -93,7 +100,7 @@ class TokenBufferMemory:
    def get_history_prompt_text(self, human_prefix: str = "Human",
                                ai_prefix: str = "Assistant",
                                max_token_limit: int = 2000,
-                                message_limit: int = 10) -> str:
+                                message_limit: Optional[int] = None) -> str:
        """
        Get history prompt text.
        :param human_prefix: human prefix
@@ -328,7 +328,7 @@ class ModelInstance:
            except Exception as e:
                raise e

-    def get_tts_voices(self, language: str) -> list:
+    def get_tts_voices(self, language: Optional[str] = None) -> list:
        """
        Invoke large language tts model voices

@@ -336,7 +336,7 @@ Inherit the `__base.text2speech_model.Text2SpeechModel` base class and implement
 - Invoke Invocation

  ```python
-  def _invoke(elf, model: str, credentials: dict, content_text: str, streaming: bool, user: Optional[str] = None):
+  def _invoke(self, model: str, credentials: dict, content_text: str, streaming: bool, user: Optional[str] = None):
      """
      Invoke large language model
  
@@ -376,7 +376,7 @@ class XinferenceProvider(Provider):
 - Invoke 调用

  ```python
-  def _invoke(elf, model: str, credentials: dict, content_text: str, streaming: bool, user: Optional[str] = None):
+  def _invoke(self, model: str, credentials: dict, content_text: str, streaming: bool, user: Optional[str] = None):
      """
      Invoke large language model
  
@@ -31,3 +31,5 @@
 - volcengine_maas
 - openai_api_compatible
 - deepseek
+- hunyuan
+- siliconflow
@@ -53,6 +53,15 @@ model_credential_schema:
      type: select
      required: true
      options:
+        - label:
+            en_US: 2024-05-01-preview
+          value: 2024-05-01-preview
+        - label:
+            en_US: 2024-04-01-preview
+          value: 2024-04-01-preview
+        - label:
+            en_US: 2024-03-01-preview
+          value: 2024-03-01-preview
        - label:
            en_US: 2024-02-15-preview
          value: 2024-02-15-preview
@@ -57,23 +57,23 @@ class BaichuanModel:
        }[model]

    def _handle_chat_generate_response(self, response) -> BaichuanMessage:
-            resp = response.json()
-            choices = resp.get('choices', [])
-            message = BaichuanMessage(content='', role='assistant')
-            for choice in choices:
-                message.content += choice['message']['content']
-                message.role = choice['message']['role']
-                if choice['finish_reason']:
-                    message.stop_reason = choice['finish_reason']
+        resp = response.json()
+        choices = resp.get('choices', [])
+        message = BaichuanMessage(content='', role='assistant')
+        for choice in choices:
+            message.content += choice['message']['content']
+            message.role = choice['message']['role']
+            if choice['finish_reason']:
+                message.stop_reason = choice['finish_reason']

-            if 'usage' in resp:
-                message.usage = {
-                    'prompt_tokens': resp['usage']['prompt_tokens'],
-                    'completion_tokens': resp['usage']['completion_tokens'],
-                    'total_tokens': resp['usage']['total_tokens'],
-                }
-            
-            return message
+        if 'usage' in resp:
+            message.usage = {
+                'prompt_tokens': resp['usage']['prompt_tokens'],
+                'completion_tokens': resp['usage']['completion_tokens'],
+                'total_tokens': resp['usage']['total_tokens'],
+            }
+
+        return message
    
    def _handle_chat_stream_generate_response(self, response) -> Generator:
        for line in response.iter_lines():
@@ -7,6 +7,7 @@ from core.model_runtime.entities.message_entities import (
    PromptMessage,
    PromptMessageTool,
    SystemPromptMessage,
+    ToolPromptMessage,
    UserPromptMessage,
 )
 from core.model_runtime.errors.invoke import (
@@ -32,20 +33,21 @@ from core.model_runtime.model_providers.baichuan.llm.baichuan_turbo_errors impor


 class BaichuanLarguageModel(LargeLanguageModel):
-    def _invoke(self, model: str, credentials: dict, 
-                prompt_messages: list[PromptMessage], model_parameters: dict, 
-                tools: list[PromptMessageTool] | None = None, stop: list[str] | None = None, 
+    def _invoke(self, model: str, credentials: dict,
+                prompt_messages: list[PromptMessage], model_parameters: dict,
+                tools: list[PromptMessageTool] | None = None, stop: list[str] | None = None,
                stream: bool = True, user: str | None = None) \
            -> LLMResult | Generator:
        return self._generate(model=model, credentials=credentials, prompt_messages=prompt_messages,
-                                model_parameters=model_parameters, tools=tools, stop=stop, stream=stream, user=user)
+                              model_parameters=model_parameters, tools=tools, stop=stop, stream=stream, user=user)

    def get_num_tokens(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],
                       tools: list[PromptMessageTool] | None = None) -> int:
        return self._num_tokens_from_messages(prompt_messages)

-    def _num_tokens_from_messages(self, messages: list[PromptMessage],) -> int:
+    def _num_tokens_from_messages(self, messages: list[PromptMessage], ) -> int:
        """Calculate num tokens for baichuan model"""
+
        def tokens(text: str):
            return BaichuanTokenizer._get_num_tokens(text)

@@ -85,9 +87,20 @@ class BaichuanLarguageModel(LargeLanguageModel):
        elif isinstance(message, SystemPromptMessage):
            message = cast(SystemPromptMessage, message)
            message_dict = {"role": "user", "content": message.content}
+        elif isinstance(message, ToolPromptMessage):
+            # copy from core/model_runtime/model_providers/anthropic/llm/llm.py
+            message = cast(ToolPromptMessage, message)
+            message_dict = {
+                "role": "user",
+                "content": [{
+                    "type": "tool_result",
+                    "tool_use_id": message.tool_call_id,
+                    "content": message.content
+                }]
+            }
        else:
            raise ValueError(f"Unknown message type {type(message)}")
-        
+
        return message_dict

    def validate_credentials(self, model: str, credentials: dict) -> None:
@@ -106,13 +119,13 @@ class BaichuanLarguageModel(LargeLanguageModel):
        except Exception as e:
            raise CredentialsValidateFailedError(f"Invalid API key: {e}")

-    def _generate(self, model: str, credentials: dict, prompt_messages: list[PromptMessage], 
-                 model_parameters: dict, tools: list[PromptMessageTool] | None = None, 
-                 stop: list[str] | None = None, stream: bool = True, user: str | None = None) \
+    def _generate(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],
+                  model_parameters: dict, tools: list[PromptMessageTool] | None = None,
+                  stop: list[str] | None = None, stream: bool = True, user: str | None = None) \
            -> LLMResult | Generator:
        if tools is not None and len(tools) > 0:
            raise InvokeBadRequestError("Baichuan model doesn't support tools")
-        
+
        instance = BaichuanModel(
            api_key=credentials['api_key'],
            secret_key=credentials.get('secret_key', '')
@@ -129,11 +142,12 @@ class BaichuanLarguageModel(LargeLanguageModel):
        ]

        # invoke model
-        response = instance.generate(model=model, stream=stream, messages=messages, parameters=model_parameters, timeout=60)
+        response = instance.generate(model=model, stream=stream, messages=messages, parameters=model_parameters,
+                                     timeout=60)

        if stream:
            return self._handle_chat_generate_stream_response(model, prompt_messages, credentials, response)
-        
+
        return self._handle_chat_generate_response(model, prompt_messages, credentials, response)

    def _handle_chat_generate_response(self, model: str,
@@ -141,7 +155,9 @@ class BaichuanLarguageModel(LargeLanguageModel):
                                       credentials: dict,
                                       response: BaichuanMessage) -> LLMResult:
        # convert baichuan message to llm result
-        usage = self._calc_response_usage(model=model, credentials=credentials, prompt_tokens=response.usage['prompt_tokens'], completion_tokens=response.usage['completion_tokens'])
+        usage = self._calc_response_usage(model=model, credentials=credentials,
+                                          prompt_tokens=response.usage['prompt_tokens'],
+                                          completion_tokens=response.usage['completion_tokens'])
        return LLMResult(
            model=model,
            prompt_messages=prompt_messages,
@@ -158,7 +174,9 @@ class BaichuanLarguageModel(LargeLanguageModel):
                                              response: Generator[BaichuanMessage, None, None]) -> Generator:
        for message in response:
            if message.usage:
-                usage = self._calc_response_usage(model=model, credentials=credentials, prompt_tokens=message.usage['prompt_tokens'], completion_tokens=message.usage['completion_tokens'])
+                usage = self._calc_response_usage(model=model, credentials=credentials,
+                                                  prompt_tokens=message.usage['prompt_tokens'],
+                                                  completion_tokens=message.usage['completion_tokens'])
                yield LLMResultChunk(
                    model=model,
                    prompt_messages=prompt_messages,
@@ -59,15 +59,15 @@ class BedrockTextEmbeddingModel(TextEmbeddingModel):
        model_prefix = model.split('.')[0]
         
        if model_prefix == "amazon" :
-           for text in texts:
-              body = {
+            for text in texts:
+                body = {
                 "inputText": text,
-              }
-              response_body = self._invoke_bedrock_embedding(model, bedrock_runtime, body)
-              embeddings.extend([response_body.get('embedding')])
-              token_usage += response_body.get('inputTextTokenCount')
-           logger.warning(f'Total Tokens: {token_usage}')
-           result = TextEmbeddingResult(
+                }
+                response_body = self._invoke_bedrock_embedding(model, bedrock_runtime, body)
+                embeddings.extend([response_body.get('embedding')])
+                token_usage += response_body.get('inputTextTokenCount')
+            logger.warning(f'Total Tokens: {token_usage}')
+            result = TextEmbeddingResult(
                model=model,
                embeddings=embeddings,
                usage=self._calc_response_usage(
@@ -75,20 +75,20 @@ class BedrockTextEmbeddingModel(TextEmbeddingModel):
                    credentials=credentials,
                    tokens=token_usage
                )
-           )
-           return result
-           
+            )
+            return result
+
        if model_prefix == "cohere" :
-           input_type = 'search_document' if len(texts) > 1 else 'search_query'
-           for text in texts:
-              body = {
+            input_type = 'search_document' if len(texts) > 1 else 'search_query'
+            for text in texts:
+                body = {
                 "texts": [text],
                 "input_type": input_type,
-              }
-              response_body = self._invoke_bedrock_embedding(model, bedrock_runtime, body)
-              embeddings.extend(response_body.get('embeddings'))
-              token_usage += len(text)
-           result = TextEmbeddingResult(
+                }
+                response_body = self._invoke_bedrock_embedding(model, bedrock_runtime, body)
+                embeddings.extend(response_body.get('embeddings'))
+                token_usage += len(text)
+            result = TextEmbeddingResult(
                model=model,
                embeddings=embeddings,
                usage=self._calc_response_usage(
@@ -96,9 +96,9 @@ class BedrockTextEmbeddingModel(TextEmbeddingModel):
                    credentials=credentials,
                    tokens=token_usage
                )
-           )
-           return result
-        
+            )
+            return result
+
        #others
        raise ValueError(f"Got unknown model prefix {model_prefix} when handling block response")

@@ -183,7 +183,7 @@ class BedrockTextEmbeddingModel(TextEmbeddingModel):
        )

        return usage
-    
+
    def _map_client_to_invoke_error(self, error_code: str, error_msg: str) -> type[InvokeError]:
        """
        Map client error to invoke error
@@ -212,9 +212,9 @@ class BedrockTextEmbeddingModel(TextEmbeddingModel):
        content_type = 'application/json'
        try:
            response = bedrock_runtime.invoke_model(
-                body=json.dumps(body), 
-                modelId=model, 
-                accept=accept, 
+                body=json.dumps(body),
+                modelId=model,
+                accept=accept,
                contentType=content_type
            )
            response_body = json.loads(response.get('body').read().decode('utf-8'))
@@ -1,18 +1,22 @@
+import base64
 import json
 import logging
+import mimetypes
 from collections.abc import Generator
-from typing import Optional, Union
+from typing import Optional, Union, cast

 import google.ai.generativelanguage as glm
 import google.api_core.exceptions as exceptions
 import google.generativeai as genai
 import google.generativeai.client as client
+import requests
 from google.generativeai.types import ContentType, GenerateContentResponse, HarmBlockThreshold, HarmCategory
 from google.generativeai.types.content_types import to_part

 from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta
 from core.model_runtime.entities.message_entities import (
    AssistantPromptMessage,
+    ImagePromptMessageContent,
    PromptMessage,
    PromptMessageContentType,
    PromptMessageTool,
@@ -204,6 +208,7 @@ class GoogleLargeLanguageModel(LargeLanguageModel):
            stream=stream,
            safety_settings=safety_settings,
            tools=self._convert_tools_to_glm_tool(tools) if tools else None,
+            request_options={"timeout": 600}
        )

        if stream:
@@ -360,11 +365,22 @@ class GoogleLargeLanguageModel(LargeLanguageModel):
                for c in message.content:
                    if c.type == PromptMessageContentType.TEXT:
                        glm_content['parts'].append(to_part(c.data))
-                    else:
-                        metadata, data = c.data.split(',', 1)
-                        mime_type = metadata.split(';', 1)[0].split(':')[1]
-                        blob = {"inline_data":{"mime_type":mime_type,"data":data}}
+                    elif c.type == PromptMessageContentType.IMAGE:
+                        message_content = cast(ImagePromptMessageContent, c)
+                        if message_content.data.startswith("data:"):
+                            metadata, base64_data = c.data.split(',', 1)
+                            mime_type = metadata.split(';', 1)[0].split(':')[1]
+                        else:
+                            # fetch image data from url
+                            try:
+                                image_content = requests.get(message_content.data).content
+                                mime_type, _ = mimetypes.guess_type(message_content.data)
+                                base64_data = base64.b64encode(image_content).decode('utf-8')
+                            except Exception as ex:
+                                raise ValueError(f"Failed to fetch image data from url {message_content.data}, {ex}")
+                        blob = {"inline_data":{"mime_type":mime_type,"data":base64_data}}
                        glm_content['parts'].append(blob)
+
            return glm_content
        elif isinstance(message, AssistantPromptMessage):
            glm_content = {
@@ -443,4 +459,4 @@ class GoogleLargeLanguageModel(LargeLanguageModel):
                exceptions.RequestRangeNotSatisfiable,
                exceptions.Cancelled,
            ]
-        }
+        }
@@ -0,0 +1,30 @@
+import logging
+
+from core.model_runtime.entities.model_entities import ModelType
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.__base.model_provider import ModelProvider
+
+logger = logging.getLogger(__name__)
+
+class HunyuanProvider(ModelProvider):
+
+    def validate_provider_credentials(self, credentials: dict) -> None:
+        """
+        Validate provider credentials
+        if validate failed, raise exception
+
+        :param credentials: provider credentials, credentials form defined in `provider_credential_schema`.
+        """
+        try:
+            model_instance = self.get_model_instance(ModelType.LLM)
+
+            # Use `hunyuan-standard` model for validate,
+            model_instance.validate_credentials(
+                model='hunyuan-standard',
+                credentials=credentials
+            )
+        except CredentialsValidateFailedError as ex:
+            raise ex
+        except Exception as ex:
+            logger.exception(f'{self.get_provider_schema().provider} credentials validate failed')
+            raise ex
@@ -0,0 +1,40 @@
+provider: hunyuan
+label:
+  zh_Hans: 腾讯混元
+  en_US: Hunyuan
+description:
+  en_US: Models provided by Tencent Hunyuan, such as hunyuan-standard, hunyuan-standard-256k, hunyuan-pro and hunyuan-lite.
+  zh_Hans: 腾讯混元提供的模型，例如 hunyuan-standard、 hunyuan-standard-256k, hunyuan-pro 和 hunyuan-lite。
+icon_small:
+  en_US: icon_s_en.png
+icon_large:
+  en_US: icon_l_en.png
+background: "#F6F7F7"
+help:
+  title:
+    en_US: Get your API Key from Tencent Hunyuan
+    zh_Hans: 从腾讯混元获取 API Key
+  url:
+    en_US: https://console.cloud.tencent.com/cam/capi
+supported_model_types:
+  - llm
+configurate_methods:
+  - predefined-model
+provider_credential_schema:
+  credential_form_schemas:
+    - variable: secret_id
+      label:
+        en_US: Secret ID
+      type: secret-input
+      required: true
+      placeholder:
+        zh_Hans: 在此输入您的 Secret ID
+        en_US: Enter your Secret ID
+    - variable: secret_key
+      label:
+        en_US: Secret Key
+      type: secret-input
+      required: true
+      placeholder:
+        zh_Hans: 在此输入您的 Secret Key
+        en_US: Enter your Secret Key
@@ -0,0 +1,4 @@
+- hunyuan-lite
+- hunyuan-standard
+- hunyuan-standard-256k
+- hunyuan-pro
@@ -0,0 +1,28 @@
+model: hunyuan-lite
+label:
+  zh_Hans: hunyuan-lite
+  en_US: hunyuan-lite
+model_type: llm
+features:
+  - agent-thought
+  - tool-call
+  - multi-tool-call
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 256000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 256000
+pricing:
+  input: '0.00'
+  output: '0.00'
+  unit: '0.001'
+  currency: RMB
@@ -0,0 +1,28 @@
+model: hunyuan-pro
+label:
+  zh_Hans: hunyuan-pro
+  en_US: hunyuan-pro
+model_type: llm
+features:
+  - agent-thought
+  - tool-call
+  - multi-tool-call
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 32000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 32000
+pricing:
+  input: '0.03'
+  output: '0.10'
+  unit: '0.001'
+  currency: RMB
@@ -0,0 +1,28 @@
+model: hunyuan-standard-256k
+label:
+  zh_Hans: hunyuan-standard-256k
+  en_US: hunyuan-standard-256k
+model_type: llm
+features:
+  - agent-thought
+  - tool-call
+  - multi-tool-call
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 256000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 256000
+pricing:
+  input: '0.015'
+  output: '0.06'
+  unit: '0.001'
+  currency: RMB
@@ -0,0 +1,28 @@
+model: hunyuan-standard
+label:
+  zh_Hans: hunyuan-standard
+  en_US: hunyuan-standard
+model_type: llm
+features:
+  - agent-thought
+  - tool-call
+  - multi-tool-call
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 32000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 32000
+pricing:
+  input: '0.0045'
+  output: '0.0005'
+  unit: '0.001'
+  currency: RMB
@@ -0,0 +1,205 @@
+import json
+import logging
+from collections.abc import Generator
+
+from tencentcloud.common import credential
+from tencentcloud.common.exception import TencentCloudSDKException
+from tencentcloud.common.profile.client_profile import ClientProfile
+from tencentcloud.common.profile.http_profile import HttpProfile
+from tencentcloud.hunyuan.v20230901 import hunyuan_client, models
+
+from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMResultChunkDelta
+from core.model_runtime.entities.message_entities import (
+    AssistantPromptMessage,
+    PromptMessage,
+    PromptMessageTool,
+    SystemPromptMessage,
+    UserPromptMessage,
+)
+from core.model_runtime.errors.invoke import InvokeError
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.__base.large_language_model import LargeLanguageModel
+
+logger = logging.getLogger(__name__)
+
+class HunyuanLargeLanguageModel(LargeLanguageModel):
+
+    def _invoke(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],
+                model_parameters: dict, tools: list[PromptMessageTool] | None = None,
+                stop: list[str] | None = None, stream: bool = True, user: str | None = None) \
+            -> LLMResult | Generator:
+
+        client = self._setup_hunyuan_client(credentials)
+        request = models.ChatCompletionsRequest()
+        messages_dict = self._convert_prompt_messages_to_dicts(prompt_messages)
+
+        custom_parameters = {
+            'Temperature': model_parameters.get('temperature', 0.0),
+            'TopP': model_parameters.get('top_p', 1.0)
+        }
+
+        params = {
+            "Model": model,
+            "Messages": messages_dict,
+            "Stream": stream,
+            **custom_parameters,
+        }
+
+        request.from_json_string(json.dumps(params))
+        response = client.ChatCompletions(request)
+
+        if stream:
+            return self._handle_stream_chat_response(model, credentials, prompt_messages, response)
+
+        return self._handle_chat_response(credentials, model, prompt_messages, response)
+
+    def validate_credentials(self, model: str, credentials: dict) -> None:
+        """
+        Validate credentials
+        """
+        try:
+            client = self._setup_hunyuan_client(credentials)
+
+            req = models.ChatCompletionsRequest()
+            params = {
+                "Model": model,
+                "Messages": [{
+                    "Role": "user",
+                    "Content": "hello"
+                }],
+                "TopP": 1,
+                "Temperature": 0,
+                "Stream": False
+            }
+            req.from_json_string(json.dumps(params))
+            client.ChatCompletions(req)
+        except Exception as e:
+            raise CredentialsValidateFailedError(f'Credentials validation failed: {e}')
+
+    def _setup_hunyuan_client(self, credentials):
+        secret_id = credentials['secret_id']
+        secret_key = credentials['secret_key']
+        cred = credential.Credential(secret_id, secret_key)
+        httpProfile = HttpProfile()
+        httpProfile.endpoint = "hunyuan.tencentcloudapi.com"
+        clientProfile = ClientProfile()
+        clientProfile.httpProfile = httpProfile
+        client = hunyuan_client.HunyuanClient(cred, "", clientProfile)
+        return client
+
+    def _convert_prompt_messages_to_dicts(self, prompt_messages: list[PromptMessage]) -> list[dict]:
+        """Convert a list of PromptMessage objects to a list of dictionaries with 'Role' and 'Content' keys."""
+        return [{"Role": message.role.value, "Content": message.content} for message in prompt_messages]
+
+    def _handle_stream_chat_response(self, model, credentials, prompt_messages, resp):
+        for index, event in enumerate(resp):
+            logging.debug("_handle_stream_chat_response, event: %s", event)
+
+            data_str = event['data']
+            data = json.loads(data_str)
+
+            choices = data.get('Choices', [])
+            if not choices:
+                continue
+            choice = choices[0]
+            delta = choice.get('Delta', {})
+            message_content = delta.get('Content', '')
+            finish_reason = choice.get('FinishReason', '')
+
+            usage = data.get('Usage', {})
+            prompt_tokens = usage.get('PromptTokens', 0)
+            completion_tokens = usage.get('CompletionTokens', 0)
+            usage = self._calc_response_usage(model, credentials, prompt_tokens, completion_tokens)
+
+            assistant_prompt_message = AssistantPromptMessage(
+                content=message_content,
+                tool_calls=[]
+            )
+
+            delta_chunk = LLMResultChunkDelta(
+                index=index,
+                role=delta.get('Role', 'assistant'),
+                message=assistant_prompt_message,
+                usage=usage,
+                finish_reason=finish_reason,
+            )
+
+            yield LLMResultChunk(
+                model=model,
+                prompt_messages=prompt_messages,
+                delta=delta_chunk,
+            )
+
+    def _handle_chat_response(self, credentials, model, prompt_messages, response):
+        usage = self._calc_response_usage(model, credentials, response.Usage.PromptTokens,
+                                          response.Usage.CompletionTokens)
+        assistant_prompt_message = PromptMessage(role="assistant")
+        assistant_prompt_message.content = response.Choices[0].Message.Content
+        result = LLMResult(
+            model=model,
+            prompt_messages=prompt_messages,
+            message=assistant_prompt_message,
+            usage=usage,
+        )
+
+        return result
+
+    def get_num_tokens(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],
+                       tools: list[PromptMessageTool] | None = None) -> int:
+        if len(prompt_messages) == 0:
+            return 0
+        prompt = self._convert_messages_to_prompt(prompt_messages)
+        return self._get_num_tokens_by_gpt2(prompt)
+
+    def _convert_messages_to_prompt(self, messages: list[PromptMessage]) -> str:
+        """
+        Format a list of messages into a full prompt for the Anthropic model
+
+        :param messages: List of PromptMessage to combine.
+        :return: Combined string with necessary human_prompt and ai_prompt tags.
+        """
+        messages = messages.copy()  # don't mutate the original list
+
+        text = "".join(
+            self._convert_one_message_to_text(message)
+            for message in messages
+        )
+
+        # trim off the trailing ' ' that might come from the "Assistant: "
+        return text.rstrip()
+
+    def _convert_one_message_to_text(self, message: PromptMessage) -> str:
+        """
+        Convert a single message to a string.
+
+        :param message: PromptMessage to convert.
+        :return: String representation of the message.
+        """
+        human_prompt = "\n\nHuman:"
+        ai_prompt = "\n\nAssistant:"
+        content = message.content
+
+        if isinstance(message, UserPromptMessage):
+            message_text = f"{human_prompt} {content}"
+        elif isinstance(message, AssistantPromptMessage):
+            message_text = f"{ai_prompt} {content}"
+        elif isinstance(message, SystemPromptMessage):
+            message_text = content
+        else:
+            raise ValueError(f"Got unknown type {message}")
+
+        return message_text
+
+    @property
+    def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]:
+        """
+        Map model invoke error to unified error
+        The key is the error type thrown to the caller
+        The value is the error type thrown by the model,
+        which needs to be converted into a unified error type for the caller.
+
+        :return: Invoke error mapping
+        """
+        return {
+            InvokeError: [TencentCloudSDKException],
+        }
@@ -0,0 +1,9 @@
+model: jina-clip-v1
+model_type: text-embedding
+model_properties:
+  context_size: 8192
+  max_chunks: 2048
+pricing:
+  input: '0.001'
+  unit: '0.001'
+  currency: USD
@@ -52,16 +52,21 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
            'Content-Type': 'application/json'
        }

+        def transform_jina_input_text(model, text):
+            if model == 'jina-clip-v1':
+                return {"text": text}
+            return text
+
        data = {
            'model': model,
-            'input': texts
+            'input': [transform_jina_input_text(model, text) for text in texts]
        }

        try:
            response = post(url, headers=headers, data=dumps(data))
        except Exception as e:
            raise InvokeConnectionError(str(e))
-        
+
        if response.status_code != 200:
            try:
                resp = response.json()
@@ -75,16 +80,19 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
                else:
                    raise InvokeBadRequestError(msg)
            except JSONDecodeError as e:
-                raise InvokeServerUnavailableError(f"Failed to convert response to json: {e} with text: {response.text}")
+                raise InvokeServerUnavailableError(
+                    f"Failed to convert response to json: {e} with text: {response.text}")

        try:
            resp = response.json()
            embeddings = resp['data']
            usage = resp['usage']
        except Exception as e:
-            raise InvokeServerUnavailableError(f"Failed to convert response to json: {e} with text: {response.text}")
+            raise InvokeServerUnavailableError(
+                f"Failed to convert response to json: {e} with text: {response.text}")

-        usage = self._calc_response_usage(model=model, credentials=credentials, tokens=usage['total_tokens'])
+        usage = self._calc_response_usage(
+            model=model, credentials=credentials, tokens=usage['total_tokens'])

        result = TextEmbeddingResult(
            model=model,
@@ -122,7 +130,8 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
        try:
            self._invoke(model=model, credentials=credentials, texts=['ping'])
        except Exception as e:
-            raise CredentialsValidateFailedError(f'Credentials validation failed: {e}')
+            raise CredentialsValidateFailedError(
+                f'Credentials validation failed: {e}')

    @property
    def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]:
@@ -144,7 +153,7 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
                InvokeBadRequestError
            ]
        }
-    
+
    def _calc_response_usage(self, model: str, credentials: dict, tokens: int) -> EmbeddingUsage:
        """
        Calculate response usage
@@ -185,7 +194,8 @@ class JinaTextEmbeddingModel(TextEmbeddingModel):
            model_type=ModelType.TEXT_EMBEDDING,
            fetch_from=FetchFrom.CUSTOMIZABLE_MODEL,
            model_properties={
-                ModelPropertyKey.CONTEXT_SIZE: int(credentials.get('context_size'))
+                ModelPropertyKey.CONTEXT_SIZE: int(
+                    credentials.get('context_size'))
            }
        )

@@ -27,6 +27,7 @@ from core.model_runtime.entities.message_entities import (
    PromptMessage,
    PromptMessageTool,
    SystemPromptMessage,
+    ToolPromptMessage,
    UserPromptMessage,
 )
 from core.model_runtime.entities.model_entities import (
@@ -50,14 +51,14 @@ from core.model_runtime.model_providers.__base.large_language_model import Large
 from core.model_runtime.utils import helper


-class LocalAILarguageModel(LargeLanguageModel):
-    def _invoke(self, model: str, credentials: dict, 
-                prompt_messages: list[PromptMessage], model_parameters: dict, 
-                tools: list[PromptMessageTool] | None = None, stop: list[str] | None = None, 
+class LocalAILanguageModel(LargeLanguageModel):
+    def _invoke(self, model: str, credentials: dict,
+                prompt_messages: list[PromptMessage], model_parameters: dict,
+                tools: list[PromptMessageTool] | None = None, stop: list[str] | None = None,
                stream: bool = True, user: str | None = None) \
            -> LLMResult | Generator:
        return self._generate(model=model, credentials=credentials, prompt_messages=prompt_messages,
-                                model_parameters=model_parameters, tools=tools, stop=stop, stream=stream, user=user)
+                              model_parameters=model_parameters, tools=tools, stop=stop, stream=stream, user=user)

    def get_num_tokens(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],
                       tools: list[PromptMessageTool] | None = None) -> int:
@@ -67,8 +68,9 @@ class LocalAILarguageModel(LargeLanguageModel):
    def _num_tokens_from_messages(self, messages: list[PromptMessage], tools: list[PromptMessageTool]) -> int:
        """
            Calculate num tokens for baichuan model
-            LocalAI does not supports 
+            LocalAI does not supports
        """
+
        def tokens(text: str):
            """
                We cloud not determine which tokenizer to use, cause the model is customized.
@@ -124,7 +126,7 @@ class LocalAILarguageModel(LargeLanguageModel):
            num_tokens += self._num_tokens_for_tools(tools)

        return num_tokens
-    
+
    def _num_tokens_for_tools(self, tools: list[PromptMessageTool]) -> int:
        """
        Calculate num tokens for tool calling
@@ -133,6 +135,7 @@ class LocalAILarguageModel(LargeLanguageModel):
        :param tools: tools for tool calling
        :return: number of tokens
        """
+
        def tokens(text: str):
            return self._get_num_tokens_by_gpt2(text)

@@ -193,7 +196,7 @@ class LocalAILarguageModel(LargeLanguageModel):
            completion_model = LLMMode.COMPLETION.value
        else:
            raise ValueError(f"Unknown completion type {credentials['completion_type']}")
-            
+
        rules = [
            ParameterRule(
                name='temperature',
@@ -227,7 +230,7 @@ class LocalAILarguageModel(LargeLanguageModel):
            )
        ]

-        model_properties = { 
+        model_properties = {
            ModelPropertyKey.MODE: completion_model,
        } if completion_model else {}

@@ -246,11 +249,11 @@ class LocalAILarguageModel(LargeLanguageModel):

        return entity

-    def _generate(self, model: str, credentials: dict, prompt_messages: list[PromptMessage], 
-                 model_parameters: dict, tools: list[PromptMessageTool] | None = None, 
-                 stop: list[str] | None = None, stream: bool = True, user: str | None = None) \
+    def _generate(self, model: str, credentials: dict, prompt_messages: list[PromptMessage],
+                  model_parameters: dict, tools: list[PromptMessageTool] | None = None,
+                  stop: list[str] | None = None, stream: bool = True, user: str | None = None) \
            -> LLMResult | Generator:
-        
+
        kwargs = self._to_client_kwargs(credentials)
        # init model client
        client = OpenAI(**kwargs)
@@ -271,7 +274,7 @@ class LocalAILarguageModel(LargeLanguageModel):
            extra_model_kwargs['functions'] = [
                helper.dump_model(tool) for tool in tools
            ]
-        
+
        if completion_type == 'chat_completion':
            result = client.chat.completions.create(
                messages=[self._convert_prompt_message_to_dict(m) for m in prompt_messages],
@@ -294,24 +297,24 @@ class LocalAILarguageModel(LargeLanguageModel):
        if stream:
            if completion_type == 'completion':
                return self._handle_completion_generate_stream_response(
-                    model=model, credentials=credentials, response=result, tools=tools, 
+                    model=model, credentials=credentials, response=result, tools=tools,
                    prompt_messages=prompt_messages
                )
            return self._handle_chat_generate_stream_response(
-                model=model, credentials=credentials, response=result, tools=tools, 
+                model=model, credentials=credentials, response=result, tools=tools,
                prompt_messages=prompt_messages
            )
-        
+
        if completion_type == 'completion':
            return self._handle_completion_generate_response(
-                model=model, credentials=credentials, response=result, 
+                model=model, credentials=credentials, response=result,
                prompt_messages=prompt_messages
            )
        return self._handle_chat_generate_response(
-            model=model, credentials=credentials, response=result, tools=tools, 
+            model=model, credentials=credentials, response=result, tools=tools,
            prompt_messages=prompt_messages
        )
-    
+
    def _to_client_kwargs(self, credentials: dict) -> dict:
        """
        Convert invoke kwargs to client kwargs
@@ -321,7 +324,7 @@ class LocalAILarguageModel(LargeLanguageModel):
        """
        if not credentials['server_url'].endswith('/'):
            credentials['server_url'] += '/'
-            
+
        client_kwargs = {
            "timeout": Timeout(315.0, read=300.0, write=10.0, connect=5.0),
            "api_key": "1",
@@ -351,9 +354,20 @@ class LocalAILarguageModel(LargeLanguageModel):
        elif isinstance(message, SystemPromptMessage):
            message = cast(SystemPromptMessage, message)
            message_dict = {"role": "system", "content": message.content}
+        elif isinstance(message, ToolPromptMessage):
+            # copy from core/model_runtime/model_providers/anthropic/llm/llm.py
+            message = cast(ToolPromptMessage, message)
+            message_dict = {
+                "role": "user",
+                "content": [{
+                    "type": "tool_result",
+                    "tool_use_id": message.tool_call_id,
+                    "content": message.content
+                }]
+            }
        else:
            raise ValueError(f"Unknown message type {type(message)}")
-        
+
        return message_dict

    def _convert_prompt_message_to_completion_prompts(self, messages: list[PromptMessage]) -> str:
@@ -373,14 +387,14 @@ class LocalAILarguageModel(LargeLanguageModel):
                prompts += f'{message.content}\n'
            else:
                raise ValueError(f"Unknown message type {type(message)}")
-        
+
        return prompts

    def _handle_completion_generate_response(self, model: str,
-                                       prompt_messages: list[PromptMessage],
-                                       credentials: dict,
-                                       response: Completion,
-                                       ) -> LLMResult:
+                                             prompt_messages: list[PromptMessage],
+                                             credentials: dict,
+                                             response: Completion,
+                                             ) -> LLMResult:
        """
        Handle llm chat response

@@ -393,7 +407,7 @@ class LocalAILarguageModel(LargeLanguageModel):
        """
        if len(response.choices) == 0:
            raise InvokeServerUnavailableError("Empty response")
-        
+
        assistant_message = response.choices[0].text

        # transform assistant message to prompt message
@@ -407,7 +421,8 @@ class LocalAILarguageModel(LargeLanguageModel):
        )
        completion_tokens = self._num_tokens_from_messages(messages=[assistant_prompt_message], tools=[])

-        usage = self._calc_response_usage(model=model, credentials=credentials, prompt_tokens=prompt_tokens, completion_tokens=completion_tokens)
+        usage = self._calc_response_usage(model=model, credentials=credentials, prompt_tokens=prompt_tokens,
+                                          completion_tokens=completion_tokens)

        response = LLMResult(
            model=model,
@@ -436,7 +451,7 @@ class LocalAILarguageModel(LargeLanguageModel):
        """
        if len(response.choices) == 0:
            raise InvokeServerUnavailableError("Empty response")
-        
+
        assistant_message = response.choices[0].message

        # convert function call to tool call
@@ -452,7 +467,8 @@ class LocalAILarguageModel(LargeLanguageModel):
        prompt_tokens = self._num_tokens_from_messages(messages=prompt_messages, tools=tools)
        completion_tokens = self._num_tokens_from_messages(messages=[assistant_prompt_message], tools=tools)

-        usage = self._calc_response_usage(model=model, credentials=credentials, prompt_tokens=prompt_tokens, completion_tokens=completion_tokens)
+        usage = self._calc_response_usage(model=model, credentials=credentials, prompt_tokens=prompt_tokens,
+                                          completion_tokens=completion_tokens)

        response = LLMResult(
            model=model,
@@ -465,10 +481,10 @@ class LocalAILarguageModel(LargeLanguageModel):
        return response

    def _handle_completion_generate_stream_response(self, model: str,
-                                              prompt_messages: list[PromptMessage],
-                                              credentials: dict,
-                                              response: Stream[Completion],
-                                              tools: list[PromptMessageTool]) -> Generator:
+                                                    prompt_messages: list[PromptMessage],
+                                                    credentials: dict,
+                                                    response: Stream[Completion],
+                                                    tools: list[PromptMessageTool]) -> Generator:
        full_response = ''

        for chunk in response:
@@ -496,9 +512,9 @@ class LocalAILarguageModel(LargeLanguageModel):

                completion_tokens = self._num_tokens_from_messages(messages=[temp_assistant_prompt_message], tools=[])

-                usage = self._calc_response_usage(model=model, credentials=credentials, 
+                usage = self._calc_response_usage(model=model, credentials=credentials,
                                                  prompt_tokens=prompt_tokens, completion_tokens=completion_tokens)
-                
+
                yield LLMResultChunk(
                    model=model,
                    prompt_messages=prompt_messages,
@@ -538,7 +554,7 @@ class LocalAILarguageModel(LargeLanguageModel):

            if delta.finish_reason is None and (delta.delta.content is None or delta.delta.content == ''):
                continue
-            
+
            # check if there is a tool call in the response
            function_calls = None
            if delta.delta.function_call:
@@ -562,9 +578,9 @@ class LocalAILarguageModel(LargeLanguageModel):
                prompt_tokens = self._num_tokens_from_messages(messages=prompt_messages, tools=tools)
                completion_tokens = self._num_tokens_from_messages(messages=[temp_assistant_prompt_message], tools=[])

-                usage = self._calc_response_usage(model=model, credentials=credentials, 
+                usage = self._calc_response_usage(model=model, credentials=credentials,
                                                  prompt_tokens=prompt_tokens, completion_tokens=completion_tokens)
-                
+
                yield LLMResultChunk(
                    model=model,
                    prompt_messages=prompt_messages,
@@ -613,7 +629,7 @@ class LocalAILarguageModel(LargeLanguageModel):
                )
                tool_calls.append(tool_call)

-        return tool_calls 
+        return tool_calls

    @property
    def _invoke_error_mapping(self) -> dict[type[InvokeError], list[type[Exception]]]:
@@ -9,7 +9,7 @@ features:
  - stream-tool-call
 model_properties:
  mode: chat
-  context_size: 4096
+  context_size: 16385
 parameter_rules:
  - name: temperature
    use_template: temperature
@@ -0,0 +1 @@
+<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" fill="none" version="1.1" width="128" height="128" viewBox="0 0 128 128"><g><g style="opacity:0;"><rect x="0" y="0" width="128" height="128" rx="0" fill="#FFFFFF" fill-opacity="1"/></g><g><path d="M100.74,12L93.2335,12C69.21260000000001,12,55.3672,27.3468,55.3672,50.8672L55.3672,54.8988C52.6011,54.1056,49.7377,53.7031,46.8601,53.7031C29.816499999999998,53.7031,16,67.5196,16,84.5632C16,101.6069,29.816499999999998,115.423,46.8601,115.423C63.9037,115.423,77.72030000000001,101.6069,77.72030000000001,84.5632C77.72030000000001,82.4902,77.51140000000001,80.4223,77.0967,78.3911L77.2197,78.3911L100.74,78.3911C106.9654,78.3681,112,73.3151,112,67.08959999999999C112,60.8642,106.9654,55.8111,100.74,55.7882L100.7362,55.7882L100.6985,55.7879L100.6606,55.7882L77.2197,55.7882L77.2195,49.8663C77.2195,40.8584,83.7252,34.352900000000005,93.2335,34.352900000000005L100.5653,34.352900000000005L100.5733,34.352900000000005L100.5812,34.352900000000005L100.74,34.352900000000005L100.74,34.352900000000005C106.8469,34.2605,111.7497,29.284,111.7497,23.1764C111.7497,17.06889,106.8469,12.0923454,100.74,12L100.74,12ZM56.0347,84.5632C56.0347,79.4962,51.9271,75.3885,46.8601,75.3885C41.793099999999995,75.3885,37.6854,79.4962,37.6854,84.5632C37.6854,89.6303,41.793099999999995,93.7378,46.8601,93.7378C51.9271,93.7378,56.0347,89.6303,56.0347,84.5632Z" fill-rule="evenodd" fill="#8358F6" fill-opacity="1"/></g></g></svg>
@@ -0,0 +1,8 @@
+- deepseek-v2-chat
+- qwen2-72b-instruct
+- qwen2-57b-a14b-instruct
+- qwen2-7b-instruct
+- yi-1.5-34b-chat
+- yi-1.5-9b-chat
+- yi-1.5-6b-chat
+- glm4-9B-chat
@@ -0,0 +1,32 @@
+model: deepseek-ai/deepseek-v2-chat
+label:
+  en_US: deepseek-ai/deepseek-v2-chat
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 32768
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '1.33'
+  output: '1.33'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,32 @@
+model: zhipuai/glm4-9B-chat
+label:
+  en_US: zhipuai/glm4-9B-chat
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 32768
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '0.6'
+  output: '0.6'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,25 @@
+from collections.abc import Generator
+from typing import Optional, Union
+
+from core.model_runtime.entities.llm_entities import LLMResult
+from core.model_runtime.entities.message_entities import PromptMessage, PromptMessageTool
+from core.model_runtime.model_providers.openai_api_compatible.llm.llm import OAIAPICompatLargeLanguageModel
+
+
+class SiliconflowLargeLanguageModel(OAIAPICompatLargeLanguageModel):
+    def _invoke(self, model: str, credentials: dict,
+                prompt_messages: list[PromptMessage], model_parameters: dict,
+                tools: Optional[list[PromptMessageTool]] = None, stop: Optional[list[str]] = None,
+                stream: bool = True, user: Optional[str] = None) \
+            -> Union[LLMResult, Generator]:
+        self._add_custom_parameters(credentials)
+        return super()._invoke(model, credentials, prompt_messages, model_parameters, tools, stop, stream)
+
+    def validate_credentials(self, model: str, credentials: dict) -> None:
+        self._add_custom_parameters(credentials)
+        super().validate_credentials(model, credentials)
+
+    @classmethod
+    def _add_custom_parameters(cls, credentials: dict) -> None:
+        credentials['mode'] = 'chat'
+        credentials['endpoint_url'] = 'https://api.siliconflow.cn/v1'
@@ -0,0 +1,32 @@
+model: alibaba/Qwen2-57B-A14B-Instruct
+label:
+  en_US: alibaba/Qwen2-57B-A14B-Instruct
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 32768
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '1.26'
+  output: '1.26'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,32 @@
+model: alibaba/Qwen2-72B-Instruct
+label:
+  en_US: alibaba/Qwen2-72B-Instruct
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 32768
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '4.13'
+  output: '4.13'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,32 @@
+model: alibaba/Qwen2-7B-Instruct
+label:
+  en_US: alibaba/Qwen2-7B-Instruct
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 32768
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '0.35'
+  output: '0.35'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,32 @@
+model: 01-ai/Yi-1.5-34B-Chat
+label:
+  en_US: 01-ai/Yi-1.5-34B-Chat
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 16384
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '1.26'
+  output: '1.26'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,32 @@
+model: 01-ai/Yi-1.5-6B-Chat
+label:
+  en_US: 01-ai/Yi-1.5-6B-Chat
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 4096
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '0.35'
+  output: '0.35'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,32 @@
+model: 01-ai/Yi-1.5-9B-Chat
+label:
+  en_US: 01-ai/Yi-1.5-9B-Chat
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+  context_size: 16384
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: max_tokens
+    use_template: max_tokens
+    type: int
+    default: 512
+    min: 1
+    max: 4096
+    help:
+      zh_Hans: 指定生成结果长度的上限。如果生成结果截断，可以调大该参数。
+      en_US: Specifies the upper limit on the length of generated results. If the generated results are truncated, you can increase this parameter.
+  - name: top_p
+    use_template: top_p
+  - name: frequency_penalty
+    use_template: frequency_penalty
+pricing:
+  input: '0.42'
+  output: '0.42'
+  unit: '0.000001'
+  currency: RMB
@@ -0,0 +1,29 @@
+import logging
+
+from core.model_runtime.entities.model_entities import ModelType
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.__base.model_provider import ModelProvider
+
+logger = logging.getLogger(__name__)
+
+class SiliconflowProvider(ModelProvider):
+
+    def validate_provider_credentials(self, credentials: dict) -> None:
+        """
+        Validate provider credentials
+        if validate failed, raise exception
+
+        :param credentials: provider credentials, credentials form defined in `provider_credential_schema`.
+        """
+        try:
+            model_instance = self.get_model_instance(ModelType.LLM)
+
+            model_instance.validate_credentials(
+                model='deepseek-ai/deepseek-v2-chat',
+                credentials=credentials
+            )
+        except CredentialsValidateFailedError as ex:
+            raise ex
+        except Exception as ex:
+            logger.exception(f'{self.get_provider_schema().provider} credentials validate failed')
+            raise ex
@@ -0,0 +1,29 @@
+provider: siliconflow
+label:
+  zh_Hans: 硅基流动
+  en_US: SiliconFlow
+icon_small:
+  en_US: siliconflow_square.svg
+icon_large:
+  en_US: siliconflow.svg
+background: "#ffecff"
+help:
+  title:
+    en_US: Get your API Key from SiliconFlow
+    zh_Hans: 从 SiliconFlow 获取 API Key
+  url:
+    en_US: https://cloud.siliconflow.cn/keys
+supported_model_types:
+  - llm
+configurate_methods:
+  - predefined-model
+provider_credential_schema:
+  credential_form_schemas:
+    - variable: api_key
+      label:
+        en_US: API Key
+      type: secret-input
+      required: true
+      placeholder:
+        zh_Hans: 在此输入您的 API Key
+        en_US: Enter your API Key
@@ -7,7 +7,9 @@ from core.model_runtime.entities.message_entities import (
    ImagePromptMessageContent,
    PromptMessage,
    PromptMessageContentType,
+    PromptMessageTool,
    SystemPromptMessage,
+    ToolPromptMessage,
    UserPromptMessage,
 )
 from core.model_runtime.model_providers.volcengine_maas.errors import wrap_error
@@ -36,10 +38,11 @@ class MaaSClient(MaasService):
        client.set_sk(sk)
        return client

-    def chat(self, params: dict, messages: list[PromptMessage], stream=False) -> Generator | dict:
+    def chat(self, params: dict, messages: list[PromptMessage], stream=False, **extra_model_kwargs) -> Generator | dict:
        req = {
            'parameters': params,
-            'messages': [self.convert_prompt_message_to_maas_message(prompt) for prompt in messages]
+            'messages': [self.convert_prompt_message_to_maas_message(prompt) for prompt in messages],
+            **extra_model_kwargs,
        }
        if not stream:
            return super().chat(
@@ -89,10 +92,22 @@ class MaaSClient(MaasService):
            message = cast(AssistantPromptMessage, message)
            message_dict = {'role': ChatRole.ASSISTANT,
                            'content': message.content}
+            if message.tool_calls:
+                message_dict['tool_calls'] = [
+                    {
+                        'name': call.function.name,
+                        'arguments': call.function.arguments
+                    } for call in message.tool_calls
+                ]
        elif isinstance(message, SystemPromptMessage):
            message = cast(SystemPromptMessage, message)
            message_dict = {'role': ChatRole.SYSTEM,
                            'content': message.content}
+        elif isinstance(message, ToolPromptMessage):
+            message = cast(ToolPromptMessage, message)
+            message_dict = {'role': ChatRole.FUNCTION,
+                            'content': message.content,
+                            'name': message.tool_call_id}
        else:
            raise ValueError(f"Got unknown PromptMessage type {message}")

@@ -106,3 +121,14 @@ class MaaSClient(MaasService):
            raise wrap_error(e)

        return resp
+
+    @staticmethod
+    def transform_tool_prompt_to_maas_config(tool: PromptMessageTool):
+        return {
+            "type": "function",
+            "function": {
+                "name": tool.name,
+                "description": tool.description,
+                "parameters": tool.parameters,
+            }
+        }
@@ -119,8 +119,15 @@ class VolcengineMaaSLargeLanguageModel(LargeLanguageModel):
        if stop:
            req_params['stop'] = stop

+        extra_model_kwargs = {}
+        
+        if tools:
+            extra_model_kwargs['tools'] = [
+                MaaSClient.transform_tool_prompt_to_maas_config(tool) for tool in tools
+            ]
+
        resp = MaaSClient.wrap_exception(
-            lambda: client.chat(req_params, prompt_messages, stream))
+            lambda: client.chat(req_params, prompt_messages, stream, **extra_model_kwargs))
        if not stream:
            return self._handle_chat_response(model, credentials, prompt_messages, resp)
        return self._handle_stream_chat_response(model, credentials, prompt_messages, resp)
@@ -156,12 +163,26 @@ class VolcengineMaaSLargeLanguageModel(LargeLanguageModel):
        choice = choices[0]
        message = choice['message']

+        # parse tool calls
+        tool_calls = []
+        if message['tool_calls']:
+            for call in message['tool_calls']:
+                tool_call = AssistantPromptMessage.ToolCall(
+                    id=call['function']['name'],
+                    type=call['type'],
+                    function=AssistantPromptMessage.ToolCall.ToolCallFunction(
+                        name=call['function']['name'],
+                        arguments=call['function']['arguments']
+                    )
+                )
+                tool_calls.append(tool_call)
+
        return LLMResult(
            model=model,
            prompt_messages=prompt_messages,
            message=AssistantPromptMessage(
                content=message['content'] if message['content'] else '',
-                tool_calls=[],
+                tool_calls=tool_calls,
            ),
            usage=self._calc_usage(model, credentials, resp['usage']),
        )
@@ -252,6 +273,10 @@ class VolcengineMaaSLargeLanguageModel(LargeLanguageModel):
        if credentials.get('context_size'):
            model_properties[ModelPropertyKey.CONTEXT_SIZE] = int(
                credentials.get('context_size', 4096))
+
+        model_features = ModelConfigs.get(
+            credentials['base_model_name'], {}).get('features', [])
+
        entity = AIModelEntity(
            model=model,
            label=I18nObject(
@@ -260,7 +285,8 @@ class VolcengineMaaSLargeLanguageModel(LargeLanguageModel):
            fetch_from=FetchFrom.CUSTOMIZABLE_MODEL,
            model_type=ModelType.LLM,
            model_properties=model_properties,
-            parameter_rules=rules
+            parameter_rules=rules,
+            features=model_features,
        )

        return entity
@@ -1,3 +1,5 @@
+from core.model_runtime.entities.model_entities import ModelFeature
+
 ModelConfigs = {
    'Doubao-pro-4k': {
        'req_params': {
@@ -7,7 +9,10 @@ ModelConfigs = {
        'model_properties': {
            'context_size': 4096,
            'mode': 'chat',
-        }
+        },
+        'features': [
+            ModelFeature.TOOL_CALL
+        ],
    },
    'Doubao-lite-4k': {
        'req_params': {
@@ -17,7 +22,10 @@ ModelConfigs = {
        'model_properties': {
            'context_size': 4096,
            'mode': 'chat',
-        }
+        },
+        'features': [
+            ModelFeature.TOOL_CALL
+        ],
    },
    'Doubao-pro-32k': {
        'req_params': {
@@ -27,7 +35,10 @@ ModelConfigs = {
        'model_properties': {
            'context_size': 32768,
            'mode': 'chat',
-        }
+        },
+        'features': [
+            ModelFeature.TOOL_CALL
+        ],
    },
    'Doubao-lite-32k': {
        'req_params': {
@@ -37,7 +48,10 @@ ModelConfigs = {
        'model_properties': {
            'context_size': 32768,
            'mode': 'chat',
-        }
+        },
+        'features': [
+            ModelFeature.TOOL_CALL
+        ],
    },
    'Doubao-pro-128k': {
        'req_params': {
@@ -47,7 +61,10 @@ ModelConfigs = {
        'model_properties': {
            'context_size': 131072,
            'mode': 'chat',
-        }
+        },
+        'features': [
+            ModelFeature.TOOL_CALL
+        ],
    },
    'Doubao-lite-128k': {
        'req_params': {
@@ -57,7 +74,10 @@ ModelConfigs = {
        'model_properties': {
            'context_size': 131072,
            'mode': 'chat',
-        }
+        },
+        'features': [
+            ModelFeature.TOOL_CALL
+        ],
    },
    'Skylark2-pro-4k': {
        'req_params': {
@@ -67,6 +87,29 @@ ModelConfigs = {
        'model_properties': {
            'context_size': 4096,
            'mode': 'chat',
-        }
+        },
+        'features': [],
    },
+    'Llama3-8B': {
+        'req_params': {
+            'max_prompt_tokens': 8192,
+            'max_new_tokens': 8192,
+        },
+        'model_properties': {
+            'context_size': 8192,
+            'mode': 'chat',
+        },
+        'features': [],
+    },
+    'Llama3-70B': {
+        'req_params': {
+            'max_prompt_tokens': 8192,
+            'max_new_tokens': 8192,
+        },
+        'model_properties': {
+            'context_size': 8192,
+            'mode': 'chat',
+        },
+        'features': [],
+    }
 }
@@ -2,7 +2,8 @@ provider: volcengine_maas
 label:
  en_US: Volcengine
 description:
-  en_US: Volcengine MaaS models.
+  en_US: Volcengine Ark models.
+  zh_Hans: 火山方舟提供的模型，例如 Doubao-pro-4k、Doubao-pro-32k 和 Doubao-pro-128k。
 icon_small:
  en_US: icon_s_en.svg
 icon_large:
@@ -12,6 +13,7 @@ background: "#F9FAFB"
 help:
  title:
    en_US: Get your Access Key and Secret Access Key from Volcengine Console
+    zh_Hans: 从火山引擎控制台获取您的 Access Key 和 Secret Access Key
  url:
    en_US: https://console.volcengine.com/iam/keymanage/
 supported_model_types:
@@ -50,7 +52,7 @@ model_credential_schema:
      required: true
      label:
        en_US: Volcengine Region
-        zh_Hans: 火山引擎地区
+        zh_Hans: 火山引擎地域
      type: text-input
      default: cn-beijing
      placeholder:
@@ -124,6 +126,18 @@ model_credential_schema:
          show_on:
            - variable: __model_type
              value: llm
+        - label:
+            en_US: Llama3-8B
+          value: Llama3-8B
+          show_on:
+            - variable: __model_type
+              value: llm
+        - label:
+            en_US: Llama3-70B
+          value: Llama3-70B
+          show_on:
+            - variable: __model_type
+              value: llm
        - label:
            en_US: Doubao-embedding
          value: Doubao-embedding
@@ -22,6 +22,9 @@ parameter_rules:
    max: 2048
  - name: presence_penalty
    use_template: presence_penalty
+    default: 1.0
+    min: 1.0
+    max: 2.0
  - name: frequency_penalty
    use_template: frequency_penalty
  - name: response_format
@@ -22,6 +22,9 @@ parameter_rules:
    max: 2048
  - name: presence_penalty
    use_template: presence_penalty
+    default: 1.0
+    min: 1.0
+    max: 2.0
  - name: frequency_penalty
    use_template: frequency_penalty
  - name: response_format
@@ -0,0 +1,39 @@
+model: glm-4-0520
+label:
+  en_US: glm-4-0520
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    default: 0.95
+    min: 0.0
+    max: 1.0
+    help:
+      zh_Hans: 采样温度，控制输出的随机性，必须为正数取值范围是：(0.0,1.0]，不能等于 0,默认值为 0.95 值越大，会使输出更随机，更具创造性；值越小，输出会更加稳定或确定建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Sampling temperature, controls the randomness of the output, must be a positive number. The value range is (0.0,1.0], which cannot be equal to 0. The default value is 0.95. The larger the value, the more random and creative the output will be; the smaller the value, The output will be more stable or certain. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: top_p
+    use_template: top_p
+    default: 0.7
+    help:
+      zh_Hans: 用温度取样的另一种方法，称为核取样取值范围是：(0.0, 1.0) 开区间，不能等于 0 或 1，默认值为 0.7 模型考虑具有 top_p 概率质量tokens的结果例如：0.1 意味着模型解码器只考虑从前 10% 的概率的候选集中取 tokens 建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Another method of temperature sampling is called kernel sampling. The value range is (0.0, 1.0) open interval, which cannot be equal to 0 or 1. The default value is 0.7. The model considers the results with top_p probability mass tokens. For example 0.1 means The model decoder only considers tokens from the candidate set with the top 10% probability. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: incremental
+    label:
+      zh_Hans: 增量返回
+      en_US: Incremental
+    type: boolean
+    help:
+      zh_Hans: SSE接口调用时，用于控制每次返回内容方式是增量还是全量，不提供此参数时默认为增量返回，true 为增量返回，false 为全量返回。
+      en_US: When the SSE interface is called, it is used to control whether the content is returned incrementally or in full. If this parameter is not provided, the default is incremental return. true means incremental return, false means full return.
+    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 8192
@@ -0,0 +1,39 @@
+model: glm-4-air
+label:
+  en_US: glm-4-air
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    default: 0.95
+    min: 0.0
+    max: 1.0
+    help:
+      zh_Hans: 采样温度，控制输出的随机性，必须为正数取值范围是：(0.0,1.0]，不能等于 0,默认值为 0.95 值越大，会使输出更随机，更具创造性；值越小，输出会更加稳定或确定建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Sampling temperature, controls the randomness of the output, must be a positive number. The value range is (0.0,1.0], which cannot be equal to 0. The default value is 0.95. The larger the value, the more random and creative the output will be; the smaller the value, The output will be more stable or certain. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: top_p
+    use_template: top_p
+    default: 0.7
+    help:
+      zh_Hans: 用温度取样的另一种方法，称为核取样取值范围是：(0.0, 1.0) 开区间，不能等于 0 或 1，默认值为 0.7 模型考虑具有 top_p 概率质量tokens的结果例如：0.1 意味着模型解码器只考虑从前 10% 的概率的候选集中取 tokens 建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Another method of temperature sampling is called kernel sampling. The value range is (0.0, 1.0) open interval, which cannot be equal to 0 or 1. The default value is 0.7. The model considers the results with top_p probability mass tokens. For example 0.1 means The model decoder only considers tokens from the candidate set with the top 10% probability. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: incremental
+    label:
+      zh_Hans: 增量返回
+      en_US: Incremental
+    type: boolean
+    help:
+      zh_Hans: SSE接口调用时，用于控制每次返回内容方式是增量还是全量，不提供此参数时默认为增量返回，true 为增量返回，false 为全量返回。
+      en_US: When the SSE interface is called, it is used to control whether the content is returned incrementally or in full. If this parameter is not provided, the default is incremental return. true means incremental return, false means full return.
+    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 8192
@@ -0,0 +1,39 @@
+model: glm-4-airx
+label:
+  en_US: glm-4-airx
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    default: 0.95
+    min: 0.0
+    max: 1.0
+    help:
+      zh_Hans: 采样温度，控制输出的随机性，必须为正数取值范围是：(0.0,1.0]，不能等于 0,默认值为 0.95 值越大，会使输出更随机，更具创造性；值越小，输出会更加稳定或确定建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Sampling temperature, controls the randomness of the output, must be a positive number. The value range is (0.0,1.0], which cannot be equal to 0. The default value is 0.95. The larger the value, the more random and creative the output will be; the smaller the value, The output will be more stable or certain. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: top_p
+    use_template: top_p
+    default: 0.7
+    help:
+      zh_Hans: 用温度取样的另一种方法，称为核取样取值范围是：(0.0, 1.0) 开区间，不能等于 0 或 1，默认值为 0.7 模型考虑具有 top_p 概率质量tokens的结果例如：0.1 意味着模型解码器只考虑从前 10% 的概率的候选集中取 tokens 建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Another method of temperature sampling is called kernel sampling. The value range is (0.0, 1.0) open interval, which cannot be equal to 0 or 1. The default value is 0.7. The model considers the results with top_p probability mass tokens. For example 0.1 means The model decoder only considers tokens from the candidate set with the top 10% probability. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: incremental
+    label:
+      zh_Hans: 增量返回
+      en_US: Incremental
+    type: boolean
+    help:
+      zh_Hans: SSE接口调用时，用于控制每次返回内容方式是增量还是全量，不提供此参数时默认为增量返回，true 为增量返回，false 为全量返回。
+      en_US: When the SSE interface is called, it is used to control whether the content is returned incrementally or in full. If this parameter is not provided, the default is incremental return. true means incremental return, false means full return.
+    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 8192
@@ -0,0 +1,39 @@
+model: glm-4-flash
+label:
+  en_US: glm-4-flash
+model_type: llm
+features:
+  - multi-tool-call
+  - agent-thought
+  - stream-tool-call
+model_properties:
+  mode: chat
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+    default: 0.95
+    min: 0.0
+    max: 1.0
+    help:
+      zh_Hans: 采样温度，控制输出的随机性，必须为正数取值范围是：(0.0,1.0]，不能等于 0,默认值为 0.95 值越大，会使输出更随机，更具创造性；值越小，输出会更加稳定或确定建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Sampling temperature, controls the randomness of the output, must be a positive number. The value range is (0.0,1.0], which cannot be equal to 0. The default value is 0.95. The larger the value, the more random and creative the output will be; the smaller the value, The output will be more stable or certain. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: top_p
+    use_template: top_p
+    default: 0.7
+    help:
+      zh_Hans: 用温度取样的另一种方法，称为核取样取值范围是：(0.0, 1.0) 开区间，不能等于 0 或 1，默认值为 0.7 模型考虑具有 top_p 概率质量tokens的结果例如：0.1 意味着模型解码器只考虑从前 10% 的概率的候选集中取 tokens 建议您根据应用场景调整 top_p 或 temperature 参数，但不要同时调整两个参数。
+      en_US: Another method of temperature sampling is called kernel sampling. The value range is (0.0, 1.0) open interval, which cannot be equal to 0 or 1. The default value is 0.7. The model considers the results with top_p probability mass tokens. For example 0.1 means The model decoder only considers tokens from the candidate set with the top 10% probability. It is recommended that you adjust the top_p or temperature parameters according to the application scenario, but do not adjust both parameters at the same time.
+  - name: incremental
+    label:
+      zh_Hans: 增量返回
+      en_US: Incremental
+    type: boolean
+    help:
+      zh_Hans: SSE接口调用时，用于控制每次返回内容方式是增量还是全量，不提供此参数时默认为增量返回，true 为增量返回，false 为全量返回。
+      en_US: When the SSE interface is called, it is used to control whether the content is returned incrementally or in full. If this parameter is not provided, the default is incremental return. true means incremental return, false means full return.
+    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    default: 1024
+    min: 1
+    max: 8192
@@ -0,0 +1,4 @@
+model: embedding-2
+model_type: text-embedding
+model_properties:
+  context_size: 512
@@ -77,7 +77,7 @@ class PromptTransform:
            max_token_limit=max_token_limit,
            message_limit=memory_config.window.size
            if (memory_config.window.enabled
-               and memory_config.window.size is not None
-               and memory_config.window.size > 0)
-            else 10
+                and memory_config.window.size is not None
+                and memory_config.window.size > 0)
+            else None
        )
@@ -0,0 +1,147 @@
+import json
+from typing import Any, Optional
+
+import chromadb
+from chromadb import QueryResult, Settings
+from flask import current_app
+from pydantic import BaseModel
+
+from core.rag.datasource.entity.embedding import Embeddings
+from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
+from core.rag.models.document import Document
+from extensions.ext_redis import redis_client
+from models.dataset import Dataset
+
+
+class ChromaConfig(BaseModel):
+    host: str
+    port: int
+    tenant: str
+    database: str
+    auth_provider: Optional[str] = None
+    auth_credentials: Optional[str] = None
+
+    def to_chroma_params(self):
+        settings = Settings(
+            # auth
+            chroma_client_auth_provider=self.auth_provider,
+            chroma_client_auth_credentials=self.auth_credentials
+        )
+
+        return {
+            'host': self.host,
+            'port': self.port,
+            'ssl': False,
+            'tenant': self.tenant,
+            'database': self.database,
+            'settings': settings,
+        }
+
+
+class ChromaVector(BaseVector):
+
+    def __init__(self, collection_name: str, config: ChromaConfig):
+        super().__init__(collection_name)
+        self._client_config = config
+        self._client = chromadb.HttpClient(**self._client_config.to_chroma_params())
+
+    def get_type(self) -> str:
+        return VectorType.CHROMA
+
+    def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
+        if texts:
+            # create collection
+            self.create_collection(self._collection_name)
+
+            self.add_texts(texts, embeddings, **kwargs)
+
+    def create_collection(self, collection_name: str):
+        lock_name = 'vector_indexing_lock_{}'.format(collection_name)
+        with redis_client.lock(lock_name, timeout=20):
+            collection_exist_cache_key = 'vector_indexing_{}'.format(self._collection_name)
+            if redis_client.get(collection_exist_cache_key):
+                return
+            self._client.get_or_create_collection(collection_name)
+            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+
+    def add_texts(self, documents: list[Document], embeddings: list[list[float]], **kwargs):
+        uuids = self._get_uuids(documents)
+        texts = [d.page_content for d in documents]
+        metadatas = [d.metadata for d in documents]
+
+        collection = self._client.get_or_create_collection(self._collection_name)
+        collection.upsert(ids=uuids, documents=texts, embeddings=embeddings, metadatas=metadatas)
+
+    def delete_by_metadata_field(self, key: str, value: str):
+        collection = self._client.get_or_create_collection(self._collection_name)
+        collection.delete(where={key: {'$eq': value}})
+
+    def delete(self):
+        self._client.delete_collection(self._collection_name)
+
+    def delete_by_ids(self, ids: list[str]) -> None:
+        collection = self._client.get_or_create_collection(self._collection_name)
+        collection.delete(ids=ids)
+
+    def text_exists(self, id: str) -> bool:
+        collection = self._client.get_or_create_collection(self._collection_name)
+        response = collection.get(ids=[id])
+        return len(response) > 0
+
+    def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
+        collection = self._client.get_or_create_collection(self._collection_name)
+        results: QueryResult = collection.query(query_embeddings=query_vector, n_results=kwargs.get("top_k", 4))
+        score_threshold = kwargs.get("score_threshold", .0) if kwargs.get('score_threshold', .0) else 0.0
+
+        ids: list[str] = results['ids'][0]
+        documents: list[str] = results['documents'][0]
+        metadatas: dict[str, Any] = results['metadatas'][0]
+        distances: list[float] = results['distances'][0]
+
+        docs = []
+        for index in range(len(ids)):
+            distance = distances[index]
+            metadata = metadatas[index]
+            if distance >= score_threshold:
+                metadata['score'] = distance
+                doc = Document(
+                    page_content=documents[index],
+                    metadata=metadata,
+                )
+                docs.append(doc)
+
+        return docs
+
+    def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
+        # chroma does not support BM25 full text searching
+        return []
+
+
+class ChromaVectorFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> BaseVector:
+        if dataset.index_struct_dict:
+            class_prefix: str = dataset.index_struct_dict['vector_store']['class_prefix']
+            collection_name = class_prefix.lower()
+        else:
+            dataset_id = dataset.id
+            collection_name = Dataset.gen_collection_name_by_id(dataset_id).lower()
+            index_struct_dict = {
+                "type": VectorType.CHROMA,
+                "vector_store": {"class_prefix": collection_name}
+            }
+            dataset.index_struct = json.dumps(index_struct_dict)
+
+        config = current_app.config
+        return ChromaVector(
+            collection_name=collection_name,
+            config=ChromaConfig(
+                host=config.get('CHROMA_HOST'),
+                port=int(config.get('CHROMA_PORT')),
+                tenant=config.get('CHROMA_TENANT', chromadb.DEFAULT_TENANT),
+                database=config.get('CHROMA_DATABASE', chromadb.DEFAULT_DATABASE),
+                auth_provider=config.get('CHROMA_AUTH_PROVIDER'),
+                auth_credentials=config.get('CHROMA_AUTH_CREDENTIALS'),
+            ),
+        )
@@ -1,14 +1,20 @@
+import json
 import logging
 from typing import Any, Optional
 from uuid import uuid4

+from flask import current_app
 from pydantic import BaseModel, root_validator
 from pymilvus import MilvusClient, MilvusException, connections

+from core.rag.datasource.entity.embedding import Embeddings
 from core.rag.datasource.vdb.field import Field
 from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.models.document import Document
 from extensions.ext_redis import redis_client
+from models.dataset import Dataset

 logger = logging.getLogger(__name__)

@@ -55,7 +61,7 @@ class MilvusVector(BaseVector):
        self._fields = []

    def get_type(self) -> str:
-        return 'milvus'
+        return VectorType.MILVUS

    def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
        index_params = {
@@ -254,10 +260,36 @@ class MilvusVector(BaseVector):
                                                           schema=schema, index_param=index_params,
                                                           consistency_level=self._consistency_level)
            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+
    def _init_client(self, config) -> MilvusClient:
        if config.secure:
            uri = "https://" + str(config.host) + ":" + str(config.port)
        else:
            uri = "http://" + str(config.host) + ":" + str(config.port)
-        client = MilvusClient(uri=uri, user=config.user, password=config.password,db_name=config.database)
+        client = MilvusClient(uri=uri, user=config.user, password=config.password, db_name=config.database)
        return client
+
+
+class MilvusVectorFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> MilvusVector:
+        if dataset.index_struct_dict:
+            class_prefix: str = dataset.index_struct_dict['vector_store']['class_prefix']
+            collection_name = class_prefix
+        else:
+            dataset_id = dataset.id
+            collection_name = Dataset.gen_collection_name_by_id(dataset_id)
+            dataset.index_struct = json.dumps(
+                self.gen_index_struct_dict(VectorType.WEAVIATE, collection_name))
+
+        config = current_app.config
+        return MilvusVector(
+            collection_name=collection_name,
+            config=MilvusConfig(
+                host=config.get('MILVUS_HOST'),
+                port=config.get('MILVUS_PORT'),
+                user=config.get('MILVUS_USER'),
+                password=config.get('MILVUS_PASSWORD'),
+                secure=config.get('MILVUS_SECURE'),
+                database=config.get('MILVUS_DATABASE'),
+            )
+        )
@@ -1,7 +1,9 @@
+import json
 import logging
 from typing import Any
 from uuid import UUID, uuid4

+from flask import current_app
 from numpy import ndarray
 from pgvecto_rs.sqlalchemy import Vector
 from pydantic import BaseModel, root_validator
@@ -10,10 +12,14 @@ from sqlalchemy import text as sql_text
 from sqlalchemy.dialects import postgresql
 from sqlalchemy.orm import Mapped, Session, mapped_column

+from core.rag.datasource.entity.embedding import Embeddings
 from core.rag.datasource.vdb.pgvecto_rs.collection import CollectionORM
 from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.models.document import Document
 from extensions.ext_redis import redis_client
+from models.dataset import Dataset

 logger = logging.getLogger(__name__)

@@ -54,7 +60,7 @@ class PGVectoRS(BaseVector):

        class _Table(CollectionORM):
            __tablename__ = collection_name
-            __table_args__ = {"extend_existing": True}  # noqa: RUF012
+            __table_args__ = {"extend_existing": True}
            id: Mapped[UUID] = mapped_column(
                postgresql.UUID(as_uuid=True),
                primary_key=True,
@@ -67,7 +73,7 @@ class PGVectoRS(BaseVector):
        self._distance_op = "<=>"

    def get_type(self) -> str:
-        return 'pgvecto-rs'
+        return VectorType.PGVECTO_RS

    def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
        self.create_collection(len(embeddings[0]))
@@ -222,3 +228,28 @@ class PGVectoRS(BaseVector):
        #         docs.append(doc)
        #     return docs
        return []
+
+
+class PGVectoRSFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> PGVectoRS:
+        if dataset.index_struct_dict:
+            class_prefix: str = dataset.index_struct_dict['vector_store']['class_prefix']
+            collection_name = class_prefix.lower()
+        else:
+            dataset_id = dataset.id
+            collection_name = Dataset.gen_collection_name_by_id(dataset_id).lower()
+            dataset.index_struct = json.dumps(
+                self.gen_index_struct_dict(VectorType.WEAVIATE, collection_name))
+        dim = len(embeddings.embed_query("pgvecto_rs"))
+        config = current_app.config
+        return PGVectoRS(
+            collection_name=collection_name,
+            config=PgvectoRSConfig(
+                host=config.get('PGVECTO_RS_HOST'),
+                port=config.get('PGVECTO_RS_PORT'),
+                user=config.get('PGVECTO_RS_USER'),
+                password=config.get('PGVECTO_RS_PASSWORD'),
+                database=config.get('PGVECTO_RS_DATABASE'),
+            ),
+            dim=dim
+        )
@@ -5,11 +5,16 @@ from typing import Any

 import psycopg2.extras
 import psycopg2.pool
+from flask import current_app
 from pydantic import BaseModel, root_validator

+from core.rag.datasource.entity.embedding import Embeddings
 from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.models.document import Document
 from extensions.ext_redis import redis_client
+from models.dataset import Dataset


 class PGVectorConfig(BaseModel):
@@ -51,7 +56,7 @@ class PGVector(BaseVector):
        self.table_name = f"embedding_{collection_name}"

    def get_type(self) -> str:
-        return "pgvector"
+        return VectorType.PGVECTOR

    def _create_connection_pool(self, config: PGVectorConfig):
        return psycopg2.pool.SimpleConnectionPool(
@@ -167,3 +172,27 @@ class PGVector(BaseVector):
                cur.execute(SQL_CREATE_TABLE.format(table_name=self.table_name, dimension=dimension))
                # TODO: create index https://github.com/pgvector/pgvector?tab=readme-ov-file#indexing
            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+
+
+class PGVectorFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> PGVector:
+        if dataset.index_struct_dict:
+            class_prefix: str = dataset.index_struct_dict["vector_store"]["class_prefix"]
+            collection_name = class_prefix
+        else:
+            dataset_id = dataset.id
+            collection_name = Dataset.gen_collection_name_by_id(dataset_id)
+            dataset.index_struct = json.dumps(
+                self.gen_index_struct_dict(VectorType.PGVECTOR, collection_name))
+
+        config = current_app.config
+        return PGVector(
+            collection_name=collection_name,
+            config=PGVectorConfig(
+                host=config.get("PGVECTOR_HOST"),
+                port=config.get("PGVECTOR_PORT"),
+                user=config.get("PGVECTOR_USER"),
+                password=config.get("PGVECTOR_PASSWORD"),
+                database=config.get("PGVECTOR_DATABASE"),
+            ),
+        )
@@ -1,3 +1,4 @@
+import json
 import os
 import uuid
 from collections.abc import Generator, Iterable, Sequence
@@ -5,6 +6,7 @@ from itertools import islice
 from typing import TYPE_CHECKING, Any, Optional, Union, cast

 import qdrant_client
+from flask import current_app
 from pydantic import BaseModel
 from qdrant_client.http import models as rest
 from qdrant_client.http.models import (
@@ -17,10 +19,15 @@ from qdrant_client.http.models import (
 )
 from qdrant_client.local.qdrant_local import QdrantLocal

+from core.rag.datasource.entity.embedding import Embeddings
 from core.rag.datasource.vdb.field import Field
 from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.models.document import Document
+from extensions.ext_database import db
 from extensions.ext_redis import redis_client
+from models.dataset import Dataset, DatasetCollectionBinding

 if TYPE_CHECKING:
    from qdrant_client import grpc  # noqa
@@ -69,7 +76,7 @@ class QdrantVector(BaseVector):
        self._group_id = group_id

    def get_type(self) -> str:
-        return 'qdrant'
+        return VectorType.QDRANT

    def to_index_struct(self) -> dict:
        return {
@@ -121,7 +128,7 @@ class QdrantVector(BaseVector):
                # create doc_id payload index
                self._client.create_payload_index(collection_name, Field.DOC_ID.value,
                                                  field_schema=PayloadSchemaType.KEYWORD)
-                # creat full text index
+                # create full text index
                text_index_params = TextIndexParams(
                    type=TextIndexType.TEXT,
                    tokenizer=TokenizerType.MULTILINGUAL,
@@ -408,3 +415,40 @@ class QdrantVector(BaseVector):
            page_content=scored_point.payload.get(content_payload_key),
            metadata=scored_point.payload.get(metadata_payload_key) or {},
        )
+
+
+class QdrantVectorFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> QdrantVector:
+        if dataset.collection_binding_id:
+            dataset_collection_binding = db.session.query(DatasetCollectionBinding). \
+                filter(DatasetCollectionBinding.id == dataset.collection_binding_id). \
+                one_or_none()
+            if dataset_collection_binding:
+                collection_name = dataset_collection_binding.collection_name
+            else:
+                raise ValueError('Dataset Collection Bindings is not exist!')
+        else:
+            if dataset.index_struct_dict:
+                class_prefix: str = dataset.index_struct_dict['vector_store']['class_prefix']
+                collection_name = class_prefix
+            else:
+                dataset_id = dataset.id
+                collection_name = Dataset.gen_collection_name_by_id(dataset_id)
+
+        if not dataset.index_struct_dict:
+            dataset.index_struct = json.dumps(
+                self.gen_index_struct_dict(VectorType.QDRANT, collection_name))
+
+        config = current_app.config
+        return QdrantVector(
+            collection_name=collection_name,
+            group_id=dataset.id,
+            config=QdrantConfig(
+                endpoint=config.get('QDRANT_URL'),
+                api_key=config.get('QDRANT_API_KEY'),
+                root_path=config.root_path,
+                timeout=config.get('QDRANT_CLIENT_TIMEOUT'),
+                grpc_port=config.get('QDRANT_GRPC_PORT'),
+                prefer_grpc=config.get('QDRANT_GRPC_ENABLED')
+            )
+        )
@@ -1,12 +1,19 @@
+import json
 import uuid
 from typing import Any, Optional

+from flask import current_app
 from pydantic import BaseModel, root_validator
 from sqlalchemy import Column, Sequence, String, Table, create_engine, insert
 from sqlalchemy import text as sql_text
 from sqlalchemy.dialects.postgresql import JSON, TEXT
 from sqlalchemy.orm import Session

+from core.rag.datasource.entity.embedding import Embeddings
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
+from models.dataset import Dataset
+
 try:
    from sqlalchemy.orm import declarative_base
 except ImportError:
@@ -53,7 +60,7 @@ class RelytVector(BaseVector):
        self._group_id = group_id

    def get_type(self) -> str:
-        return 'relyt'
+        return VectorType.RELYT

    def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
        index_params = {}
@@ -190,7 +197,7 @@ class RelytVector(BaseVector):
                    conn.execute(chunks_table.delete().where(delete_condition))
                    return True
        except Exception as e:
-            print("Delete operation failed:", str(e))  # noqa: T201
+            print("Delete operation failed:", str(e))
            return False

    def delete_by_metadata_field(self, key: str, value: str):
@@ -240,10 +247,10 @@ class RelytVector(BaseVector):
        return docs

    def similarity_search_with_score_by_vector(
-        self,
-        embedding: list[float],
-        k: int = 4,
-        filter: Optional[dict] = None,
+            self,
+            embedding: list[float],
+            k: int = 4,
+            filter: Optional[dict] = None,
    ) -> list[tuple[Document, float]]:
        # Add the filter if provided
        try:
@@ -298,3 +305,28 @@ class RelytVector(BaseVector):
    def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
        # milvus/zilliz/relyt doesn't support bm25 search
        return []
+
+
+class RelytVectorFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> RelytVector:
+        if dataset.index_struct_dict:
+            class_prefix: str = dataset.index_struct_dict['vector_store']['class_prefix']
+            collection_name = class_prefix
+        else:
+            dataset_id = dataset.id
+            collection_name = Dataset.gen_collection_name_by_id(dataset_id)
+            dataset.index_struct = json.dumps(
+                self.gen_index_struct_dict(VectorType.RELYT, collection_name))
+
+        config = current_app.config
+        return RelytVector(
+            collection_name=collection_name,
+            config=RelytConfig(
+                host=config.get('RELYT_HOST'),
+                port=config.get('RELYT_PORT'),
+                user=config.get('RELYT_USER'),
+                password=config.get('RELYT_PASSWORD'),
+                database=config.get('RELYT_DATABASE'),
+            ),
+            group_id=dataset.id
+        )
@@ -0,0 +1,249 @@
+import json
+import logging
+from typing import Any
+
+import sqlalchemy
+from flask import current_app
+from pydantic import BaseModel, root_validator
+from sqlalchemy import JSON, TEXT, Column, DateTime, String, Table, create_engine, insert
+from sqlalchemy import text as sql_text
+from sqlalchemy.orm import Session, declarative_base
+
+from core.rag.datasource.entity.embedding import Embeddings
+from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
+from core.rag.models.document import Document
+from extensions.ext_redis import redis_client
+from models.dataset import Dataset
+
+logger = logging.getLogger(__name__)
+
+
+class TiDBVectorConfig(BaseModel):
+    host: str
+    port: int
+    user: str
+    password: str
+    database: str
+
+    @root_validator()
+    def validate_config(cls, values: dict) -> dict:
+        if not values['host']:
+            raise ValueError("config TIDB_VECTOR_HOST is required")
+        if not values['port']:
+            raise ValueError("config TIDB_VECTOR_PORT is required")
+        if not values['user']:
+            raise ValueError("config TIDB_VECTOR_USER is required")
+        if not values['password']:
+            raise ValueError("config TIDB_VECTOR_PASSWORD is required")
+        if not values['database']:
+            raise ValueError("config TIDB_VECTOR_DATABASE is required")
+        return values
+
+
+class TiDBVector(BaseVector):
+
+    def get_type(self) -> str:
+        return VectorType.TIDB_VECTOR
+
+    def _table(self, dim: int) -> Table:
+        from tidb_vector.sqlalchemy import VectorType
+        return Table(
+            self._collection_name,
+            self._orm_base.metadata,
+            Column('id', String(36), primary_key=True, nullable=False),
+            Column("vector", VectorType(dim), nullable=False, comment="" if self._distance_func is None else f"hnsw(distance={self._distance_func})"),
+            Column("text", TEXT, nullable=False),
+            Column("meta", JSON, nullable=False),
+            Column("create_time", DateTime, server_default=sqlalchemy.text("CURRENT_TIMESTAMP")),
+            Column("update_time", DateTime, server_default=sqlalchemy.text("CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP")),
+            extend_existing=True
+        )
+
+    def __init__(self, collection_name: str, config: TiDBVectorConfig, distance_func: str = 'cosine'):
+        super().__init__(collection_name)
+        self._client_config = config
+        self._url = (f"mysql+pymysql://{config.user}:{config.password}@{config.host}:{config.port}/{config.database}?"
+                     f"ssl_verify_cert=true&ssl_verify_identity=true")
+        self._distance_func = distance_func.lower()
+        self._engine = create_engine(self._url)
+        self._orm_base = declarative_base()
+        self._dimension = 1536
+
+    def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
+        logger.info("create collection and add texts, collection_name: " + self._collection_name)
+        self._create_collection(len(embeddings[0]))
+        self.add_texts(texts, embeddings)
+        self._dimension = len(embeddings[0])
+        pass
+
+    def _create_collection(self, dimension: int):
+        logger.info("_create_collection, collection_name " + self._collection_name)
+        lock_name = 'vector_indexing_lock_{}'.format(self._collection_name)
+        with redis_client.lock(lock_name, timeout=20):
+            collection_exist_cache_key = 'vector_indexing_{}'.format(self._collection_name)
+            if redis_client.get(collection_exist_cache_key):
+                return
+            with Session(self._engine) as session:
+                session.begin()
+                create_statement = sql_text(f"""
+                    CREATE TABLE IF NOT EXISTS {self._collection_name} (
+                        id CHAR(36) PRIMARY KEY,
+                        text TEXT NOT NULL,
+                        meta JSON NOT NULL,
+                        doc_id VARCHAR(64) AS (JSON_UNQUOTE(JSON_EXTRACT(meta, '$.doc_id'))) STORED,
+                        KEY (doc_id),
+                        vector VECTOR<FLOAT>({dimension}) NOT NULL COMMENT "hnsw(distance={self._distance_func})",
+                        create_time DATETIME DEFAULT CURRENT_TIMESTAMP,
+                        update_time DATETIME DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP
+                    );
+                """)
+                session.execute(create_statement)
+                # tidb vector not support 'CREATE/ADD INDEX' now
+                session.commit()
+            redis_client.set(collection_exist_cache_key, 1, ex=3600)
+
+    def add_texts(self, documents: list[Document], embeddings: list[list[float]], **kwargs):
+        table = self._table(len(embeddings[0]))
+        ids = self._get_uuids(documents)
+        metas = [d.metadata for d in documents]
+        texts = [d.page_content for d in documents]
+
+        chunks_table_data = []
+        with self._engine.connect() as conn:
+            with conn.begin():
+                for id, text, meta, embedding in zip(
+                        ids, texts, metas, embeddings
+                ):
+                    chunks_table_data.append({"id": id, "vector": embedding, "text": text, "meta": meta})
+
+                    # Execute the batch insert when the batch size is reached
+                    if len(chunks_table_data) == 500:
+                        conn.execute(insert(table).values(chunks_table_data))
+                        # Clear the chunks_table_data list for the next batch
+                        chunks_table_data.clear()
+
+                # Insert any remaining records that didn't make up a full batch
+                if chunks_table_data:
+                    conn.execute(insert(table).values(chunks_table_data))
+        return ids
+
+    def text_exists(self, id: str) -> bool:
+        result = self.get_ids_by_metadata_field('doc_id', id)
+        return bool(result)
+
+    def delete_by_ids(self, ids: list[str]) -> None:
+        with Session(self._engine) as session:
+            ids_str = ','.join(f"'{doc_id}'" for doc_id in ids)
+            select_statement = sql_text(
+                f"""SELECT id FROM {self._collection_name} WHERE meta->>'$.doc_id' in ({ids_str}); """
+            )
+            result = session.execute(select_statement).fetchall()
+        if result:
+            ids = [item[0] for item in result]
+            self._delete_by_ids(ids)
+
+    def _delete_by_ids(self, ids: list[str]) -> bool:
+        if ids is None:
+            raise ValueError("No ids provided to delete.")
+        table = self._table(self._dimension)
+        try:
+            with self._engine.connect() as conn:
+                with conn.begin():
+                    delete_condition = table.c.id.in_(ids)
+                    conn.execute(table.delete().where(delete_condition))
+                    return True
+        except Exception as e:
+            print("Delete operation failed:", str(e))
+            return False
+
+    def delete_by_document_id(self, document_id: str):
+        ids = self.get_ids_by_metadata_field('document_id', document_id)
+        if ids:
+            self._delete_by_ids(ids)
+
+    def get_ids_by_metadata_field(self, key: str, value: str):
+        with Session(self._engine) as session:
+            select_statement = sql_text(
+                f"""SELECT id FROM {self._collection_name} WHERE meta->>'$.{key}' = '{value}'; """
+            )
+            result = session.execute(select_statement).fetchall()
+        if result:
+            return [item[0] for item in result]
+        else:
+            return None
+
+    def delete_by_metadata_field(self, key: str, value: str) -> None:
+        ids = self.get_ids_by_metadata_field(key, value)
+        if ids:
+            self._delete_by_ids(ids)
+
+    def search_by_vector(self, query_vector: list[float], **kwargs: Any) -> list[Document]:
+        top_k = kwargs.get("top_k", 5)
+        score_threshold = kwargs.get("score_threshold") if kwargs.get("score_threshold") else 0.0
+        filter = kwargs.get('filter')
+        distance = 1 - score_threshold
+
+        query_vector_str = ", ".join(format(x) for x in query_vector)
+        query_vector_str = "[" + query_vector_str + "]"
+        logger.debug(f"_collection_name: {self._collection_name}, score_threshold: {score_threshold}, distance: {distance}")
+
+        docs = []
+        if self._distance_func == 'l2':
+            tidb_func = 'Vec_l2_distance'
+        elif self._distance_func == 'cosine':
+            tidb_func = 'Vec_Cosine_distance'
+        else:
+            tidb_func = 'Vec_Cosine_distance'
+
+        with Session(self._engine) as session:
+            select_statement = sql_text(
+                f"""SELECT meta, text, distance FROM (
+                        SELECT meta, text, {tidb_func}(vector, "{query_vector_str}")  as distance 
+                        FROM {self._collection_name} 
+                        ORDER BY distance
+                        LIMIT {top_k}
+                    ) t WHERE distance < {distance};"""
+            )
+            res = session.execute(select_statement)
+            results = [(row[0], row[1], row[2]) for row in res]
+            for meta, text, distance in results:
+                metadata = json.loads(meta)
+                metadata['score'] = 1 - distance
+                docs.append(Document(page_content=text, metadata=metadata))
+        return docs
+
+    def search_by_full_text(self, query: str, **kwargs: Any) -> list[Document]:
+        # tidb doesn't support bm25 search
+        return []
+
+    def delete(self) -> None:
+        with Session(self._engine) as session:
+            session.execute(sql_text(f"""DROP TABLE IF EXISTS {self._collection_name};"""))
+            session.commit()
+
+
+class TiDBVectorFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> TiDBVector:
+
+        if dataset.index_struct_dict:
+            class_prefix: str = dataset.index_struct_dict['vector_store']['class_prefix']
+            collection_name = class_prefix.lower()
+        else:
+            dataset_id = dataset.id
+            collection_name = Dataset.gen_collection_name_by_id(dataset_id).lower()
+            dataset.index_struct = json.dumps(
+                self.gen_index_struct_dict(VectorType.TIDB_VECTOR, collection_name))
+
+        config = current_app.config
+        return TiDBVector(
+            collection_name=collection_name,
+            config=TiDBVectorConfig(
+                host=config.get('TIDB_VECTOR_HOST'),
+                port=config.get('TIDB_VECTOR_PORT'),
+                user=config.get('TIDB_VECTOR_USER'),
+                password=config.get('TIDB_VECTOR_PASSWORD'),
+                database=config.get('TIDB_VECTOR_DATABASE'),
+            ),
+        )
@@ -11,6 +11,10 @@ class BaseVector(ABC):
    def __init__(self, collection_name: str):
        self._collection_name = collection_name

+    @abstractmethod
+    def get_type(self) -> str:
+        raise NotImplementedError
+
    @abstractmethod
    def create(self, texts: list[Document], embeddings: list[list[float]], **kwargs):
        raise NotImplementedError
@@ -1,4 +1,4 @@
-import json
+from abc import ABC, abstractmethod
 from typing import Any

 from flask import current_app
@@ -8,9 +8,23 @@ from core.model_manager import ModelManager
 from core.model_runtime.entities.model_entities import ModelType
 from core.rag.datasource.entity.embedding import Embeddings
 from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.models.document import Document
-from extensions.ext_database import db
-from models.dataset import Dataset, DatasetCollectionBinding
+from models.dataset import Dataset
+
+
+class AbstractVectorFactory(ABC):
+    @abstractmethod
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> BaseVector:
+        raise NotImplementedError
+
+    @staticmethod
+    def gen_index_struct_dict(vector_type: VectorType, collection_name: str) -> dict:
+        index_struct_dict = {
+            "type": vector_type,
+            "vector_store": {"class_prefix": collection_name}
+        }
+        return index_struct_dict


 class Vector:
@@ -32,163 +46,38 @@ class Vector:
        if not vector_type:
            raise ValueError("Vector store must be specified.")

-        if vector_type == "weaviate":
-            from core.rag.datasource.vdb.weaviate.weaviate_vector import WeaviateConfig, WeaviateVector
-            if self._dataset.index_struct_dict:
-                class_prefix: str = self._dataset.index_struct_dict['vector_store']['class_prefix']
-                collection_name = class_prefix
-            else:
-                dataset_id = self._dataset.id
-                collection_name = Dataset.gen_collection_name_by_id(dataset_id)
-                index_struct_dict = {
-                    "type": 'weaviate',
-                    "vector_store": {"class_prefix": collection_name}
-                }
-                self._dataset.index_struct = json.dumps(index_struct_dict)
-            return WeaviateVector(
-                collection_name=collection_name,
-                config=WeaviateConfig(
-                    endpoint=config.get('WEAVIATE_ENDPOINT'),
-                    api_key=config.get('WEAVIATE_API_KEY'),
-                    batch_size=int(config.get('WEAVIATE_BATCH_SIZE'))
-                ),
-                attributes=self._attributes
-            )
-        elif vector_type == "qdrant":
-            from core.rag.datasource.vdb.qdrant.qdrant_vector import QdrantConfig, QdrantVector
-            if self._dataset.collection_binding_id:
-                dataset_collection_binding = db.session.query(DatasetCollectionBinding). \
-                    filter(DatasetCollectionBinding.id == self._dataset.collection_binding_id). \
-                    one_or_none()
-                if dataset_collection_binding:
-                    collection_name = dataset_collection_binding.collection_name
-                else:
-                    raise ValueError('Dataset Collection Bindings is not exist!')
-            else:
-                if self._dataset.index_struct_dict:
-                    class_prefix: str = self._dataset.index_struct_dict['vector_store']['class_prefix']
-                    collection_name = class_prefix
-                else:
-                    dataset_id = self._dataset.id
-                    collection_name = Dataset.gen_collection_name_by_id(dataset_id)
+        vector_factory_cls = self.get_vector_factory(vector_type)
+        return vector_factory_cls().init_vector(self._dataset, self._attributes, self._embeddings)

-            if not self._dataset.index_struct_dict:
-                index_struct_dict = {
-                    "type": 'qdrant',
-                    "vector_store": {"class_prefix": collection_name}
-                }
-                self._dataset.index_struct = json.dumps(index_struct_dict)
-
-            return QdrantVector(
-                collection_name=collection_name,
-                group_id=self._dataset.id,
-                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')
-                )
-            )
-        elif vector_type == "milvus":
-            from core.rag.datasource.vdb.milvus.milvus_vector import MilvusConfig, MilvusVector
-            if self._dataset.index_struct_dict:
-                class_prefix: str = self._dataset.index_struct_dict['vector_store']['class_prefix']
-                collection_name = class_prefix
-            else:
-                dataset_id = self._dataset.id
-                collection_name = Dataset.gen_collection_name_by_id(dataset_id)
-                index_struct_dict = {
-                    "type": 'milvus',
-                    "vector_store": {"class_prefix": collection_name}
-                }
-                self._dataset.index_struct = json.dumps(index_struct_dict)
-            return MilvusVector(
-                collection_name=collection_name,
-                config=MilvusConfig(
-                    host=config.get('MILVUS_HOST'),
-                    port=config.get('MILVUS_PORT'),
-                    user=config.get('MILVUS_USER'),
-                    password=config.get('MILVUS_PASSWORD'),
-                    secure=config.get('MILVUS_SECURE'),
-                    database=config.get('MILVUS_DATABASE'),
-                )
-            )
-        elif vector_type == "relyt":
-            from core.rag.datasource.vdb.relyt.relyt_vector import RelytConfig, RelytVector
-            if self._dataset.index_struct_dict:
-                class_prefix: str = self._dataset.index_struct_dict['vector_store']['class_prefix']
-                collection_name = class_prefix
-            else:
-                dataset_id = self._dataset.id
-                collection_name = Dataset.gen_collection_name_by_id(dataset_id)
-                index_struct_dict = {
-                    "type": 'relyt',
-                    "vector_store": {"class_prefix": collection_name}
-                }
-                self._dataset.index_struct = json.dumps(index_struct_dict)
-            return RelytVector(
-                collection_name=collection_name,
-                config=RelytConfig(
-                    host=config.get('RELYT_HOST'),
-                    port=config.get('RELYT_PORT'),
-                    user=config.get('RELYT_USER'),
-                    password=config.get('RELYT_PASSWORD'),
-                    database=config.get('RELYT_DATABASE'),
-                ),
-                group_id=self._dataset.id
-            )
-        elif vector_type == "pgvecto_rs":
-            from core.rag.datasource.vdb.pgvecto_rs.pgvecto_rs import PGVectoRS, PgvectoRSConfig
-            if self._dataset.index_struct_dict:
-                class_prefix: str = self._dataset.index_struct_dict['vector_store']['class_prefix']
-                collection_name = class_prefix.lower()
-            else:
-                dataset_id = self._dataset.id
-                collection_name = Dataset.gen_collection_name_by_id(dataset_id).lower()
-                index_struct_dict = {
-                    "type": 'pgvecto_rs',
-                    "vector_store": {"class_prefix": collection_name}
-                }
-                self._dataset.index_struct = json.dumps(index_struct_dict)
-            dim = len(self._embeddings.embed_query("pgvecto_rs"))
-            return PGVectoRS(
-                collection_name=collection_name,
-                config=PgvectoRSConfig(
-                    host=config.get('PGVECTO_RS_HOST'),
-                    port=config.get('PGVECTO_RS_PORT'),
-                    user=config.get('PGVECTO_RS_USER'),
-                    password=config.get('PGVECTO_RS_PASSWORD'),
-                    database=config.get('PGVECTO_RS_DATABASE'),
-                ),
-                dim=dim
-            )
-        elif vector_type == "pgvector":
-            from core.rag.datasource.vdb.pgvector.pgvector import PGVector, PGVectorConfig
-
-            if self._dataset.index_struct_dict:
-                class_prefix: str = self._dataset.index_struct_dict["vector_store"]["class_prefix"]
-                collection_name = class_prefix
-            else:
-                dataset_id = self._dataset.id
-                collection_name = Dataset.gen_collection_name_by_id(dataset_id)
-                index_struct_dict = {
-                    "type": "pgvector",
-                    "vector_store": {"class_prefix": collection_name}}
-                self._dataset.index_struct = json.dumps(index_struct_dict)
-            return PGVector(
-                collection_name=collection_name,
-                config=PGVectorConfig(
-                    host=config.get("PGVECTOR_HOST"),
-                    port=config.get("PGVECTOR_PORT"),
-                    user=config.get("PGVECTOR_USER"),
-                    password=config.get("PGVECTOR_PASSWORD"),
-                    database=config.get("PGVECTOR_DATABASE"),
-                ),
-            )
-        else:
-            raise ValueError(f"Vector store {config.get('VECTOR_STORE')} is not supported.")
+    @staticmethod
+    def get_vector_factory(vector_type: str) -> type[AbstractVectorFactory]:
+        match vector_type:
+            case VectorType.CHROMA:
+                from core.rag.datasource.vdb.chroma.chroma_vector import ChromaVectorFactory
+                return ChromaVectorFactory
+            case VectorType.MILVUS:
+                from core.rag.datasource.vdb.milvus.milvus_vector import MilvusVectorFactory
+                return MilvusVectorFactory
+            case VectorType.PGVECTOR:
+                from core.rag.datasource.vdb.pgvector.pgvector import PGVectorFactory
+                return PGVectorFactory
+            case VectorType.PGVECTO_RS:
+                from core.rag.datasource.vdb.pgvecto_rs.pgvecto_rs import PGVectoRSFactory
+                return PGVectoRSFactory
+            case VectorType.QDRANT:
+                from core.rag.datasource.vdb.qdrant.qdrant_vector import QdrantVectorFactory
+                return QdrantVectorFactory
+            case VectorType.RELYT:
+                from core.rag.datasource.vdb.relyt.relyt_vector import RelytVectorFactory
+                return RelytVectorFactory
+            case VectorType.TIDB_VECTOR:
+                from core.rag.datasource.vdb.tidb_vector.tidb_vector import TiDBVectorFactory
+                return TiDBVectorFactory
+            case VectorType.WEAVIATE:
+                from core.rag.datasource.vdb.weaviate.weaviate_vector import WeaviateVectorFactory
+                return WeaviateVectorFactory
+            case _:
+                raise ValueError(f"Vector store {vector_type} is not supported.")

    def create(self, texts: list = None, **kwargs):
        if texts:
@@ -0,0 +1,12 @@
+from enum import Enum
+
+
+class VectorType(str, Enum):
+    CHROMA = 'chroma'
+    MILVUS = 'milvus'
+    PGVECTOR = 'pgvector'
+    PGVECTO_RS = 'pgvecto-rs'
+    QDRANT = 'qdrant'
+    RELYT = 'relyt'
+    TIDB_VECTOR = 'tidb_vector'
+    WEAVIATE = 'weaviate'
@@ -1,12 +1,17 @@
 import datetime
+import json
 from typing import Any, Optional

 import requests
 import weaviate
+from flask import current_app
 from pydantic import BaseModel, root_validator

+from core.rag.datasource.entity.embedding import Embeddings
 from core.rag.datasource.vdb.field import Field
 from core.rag.datasource.vdb.vector_base import BaseVector
+from core.rag.datasource.vdb.vector_factory import AbstractVectorFactory
+from core.rag.datasource.vdb.vector_type import VectorType
 from core.rag.models.document import Document
 from extensions.ext_redis import redis_client
 from models.dataset import Dataset
@@ -59,7 +64,7 @@ class WeaviateVector(BaseVector):
        return client

    def get_type(self) -> str:
-        return 'weaviate'
+        return VectorType.WEAVIATE

    def get_collection_name(self, dataset: Dataset) -> str:
        if dataset.index_struct_dict:
@@ -255,3 +260,25 @@ class WeaviateVector(BaseVector):
        if isinstance(value, datetime.datetime):
            return value.isoformat()
        return value
+
+
+class WeaviateVectorFactory(AbstractVectorFactory):
+    def init_vector(self, dataset: Dataset, attributes: list, embeddings: Embeddings) -> WeaviateVector:
+        if dataset.index_struct_dict:
+            class_prefix: str = dataset.index_struct_dict['vector_store']['class_prefix']
+            collection_name = class_prefix
+        else:
+            dataset_id = dataset.id
+            collection_name = Dataset.gen_collection_name_by_id(dataset_id)
+            dataset.index_struct = json.dumps(
+                self.gen_index_struct_dict(VectorType.WEAVIATE, collection_name))
+
+        return WeaviateVector(
+            collection_name=collection_name,
+            config=WeaviateConfig(
+                endpoint=current_app.config.get('WEAVIATE_ENDPOINT'),
+                api_key=current_app.config.get('WEAVIATE_API_KEY'),
+                batch_size=int(current_app.config.get('WEAVIATE_BATCH_SIZE'))
+            ),
+            attributes=attributes
+        )
@@ -39,8 +39,8 @@ class ExcelExtractor(BaseExtractor):
        documents = []
        # loop over all sheets
        for sheet in wb.sheets():
-            for row_index, row in enumerate(sheet.get_rows(), start=1):
-                row_header = None
+            row_header = None
+            for row_index, row in enumerate(sheet.get_rows(), start=1):                
                if self.is_blank_row(row):
                    continue
                if row_header is None:
@@ -49,8 +49,8 @@ class ExcelExtractor(BaseExtractor):
                item_arr = []
                for index, cell in enumerate(row):
                    txt_value = str(cell.value)
-                    item_arr.append(f'{row_header[index].value}:{txt_value}')
-                item_str = "\n".join(item_arr)
+                    item_arr.append(f'"{row_header[index].value}":"{txt_value}"')
+                item_str = ",".join(item_arr)
                document = Document(page_content=item_str, metadata={'source': self._file_path})
                documents.append(document)
        return documents
@@ -68,7 +68,7 @@ class ExcelExtractor(BaseExtractor):

            # transform each row into a Document
            for _, row in df.iterrows():
-                item = ';'.join(f'{k}:{v}' for k, v in row.items() if pd.notna(v))
+                item = ';'.join(f'"{k}":"{v}"' for k, v in row.items() if pd.notna(v))
                document = Document(page_content=item, metadata={'source': self._file_path})
                data.append(document)
        return data
@@ -76,13 +76,25 @@ class WordExtractor(BaseExtractor):
        for rel in doc.part.rels.values():
            if "image" in rel.target_ref:
                image_count += 1
-                image_ext = rel.target_ref.split('.')[-1]
-                # user uuid as file name
-                file_uuid = str(uuid.uuid4())
-                file_key = 'image_files/' + self.tenant_id + '/' + file_uuid + '.' + image_ext
-                mime_type, _ = mimetypes.guess_type(file_key)
+                if rel.is_external:
+                    url = rel.reltype
+                    response = requests.get(url, stream=True)
+                    if response.status_code == 200:
+                        image_ext = mimetypes.guess_extension(response.headers['Content-Type'])
+                        file_uuid = str(uuid.uuid4())
+                        file_key = 'image_files/' + self.tenant_id + '/' + file_uuid + '.' + image_ext
+                        mime_type, _ = mimetypes.guess_type(file_key)
+                        storage.save(file_key, response.content)
+                    else:
+                        continue
+                else:
+                    image_ext = rel.target_ref.split('.')[-1]
+                    # user uuid as file name
+                    file_uuid = str(uuid.uuid4())
+                    file_key = 'image_files/' + self.tenant_id + '/' + file_uuid + '.' + image_ext
+                    mime_type, _ = mimetypes.guess_type(file_key)

-                storage.save(file_key, rel.target_part.blob)
+                    storage.save(file_key, rel.target_part.blob)
                # save file to db
                config = current_app.config
                upload_file = UploadFile(
@@ -50,7 +50,7 @@ class BaseDocumentTransformer(ABC):
                ) -> Sequence[Document]:
                    raise NotImplementedError

-    """  # noqa: E501
+    """

    @abstractmethod
    def transform_documents(
@@ -329,6 +329,7 @@ class DatasetRetrieval:
        """
        if not query:
            return
+        dataset_queries = []
        for dataset_id in dataset_ids:
            dataset_query = DatasetQuery(
                dataset_id=dataset_id,
@@ -338,7 +339,9 @@ class DatasetRetrieval:
                created_by_role=user_from,
                created_by=user_id
            )
-            db.session.add(dataset_query)
+            dataset_queries.append(dataset_query)
+        if dataset_queries:
+            db.session.add_all(dataset_queries)
        db.session.commit()

    def _retriever(self, flask_app: Flask, dataset_id: str, query: str, top_k: int, all_documents: list):
@@ -68,7 +68,7 @@ class ArxivAPIWrapper(BaseModel):

        Args:
            query: a plaintext search query
-        """  # noqa: E501
+        """
        try:
            results = self.arxiv_search(  # type: ignore
                query[: self.ARXIV_MAX_QUERY_LENGTH], max_results=self.top_k_results
@@ -8,10 +8,10 @@ from core.tools.tool.builtin_tool import BuiltinTool


 class DallE3Tool(BuiltinTool):
-    def _invoke(self, 
-                user_id: str, 
-               tool_parameters: dict[str, Any], 
-        ) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]:
+    def _invoke(self,
+                user_id: str,
+                tool_parameters: dict[str, Any],
+                ) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]:
        """
            invoke tools
        """
@@ -43,14 +43,18 @@ class DallE3Tool(BuiltinTool):
        style = tool_parameters.get('style', 'vivid')
        if style not in ['natural', 'vivid']:
            return self.create_text_message('Invalid style')
+        # set extra body
+        seed_id = tool_parameters.get('seed_id', self._generate_random_id(8))
+        extra_body = {'seed': seed_id}

        # call openapi dalle3
-        model=self.runtime.credentials['azure_openai_api_model_name']
+        model = self.runtime.credentials['azure_openai_api_model_name']
        response = client.images.generate(
            prompt=prompt,
            model=model,
            size=size,
            n=n,
+            extra_body=extra_body,
            style=style,
            quality=quality,
            response_format='b64_json'
@@ -59,8 +63,15 @@ class DallE3Tool(BuiltinTool):
        result = []

        for image in response.data:
-            result.append(self.create_blob_message(blob=b64decode(image.b64_json), 
-                                                   meta={ 'mime_type': 'image/png' },
-                                                    save_as=self.VARIABLE_KEY.IMAGE.value))
+            result.append(self.create_blob_message(blob=b64decode(image.b64_json),
+                                                   meta={'mime_type': 'image/png'},
+                                                   save_as=self.VARIABLE_KEY.IMAGE.value))
+        result.append(self.create_text_message(f'\nGenerate image source to Seed ID: {seed_id}'))

        return result
+
+    @staticmethod
+    def _generate_random_id(length=8):
+        characters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789'
+        random_id = ''.join(random.choices(characters, k=length))
+        return random_id
@@ -29,6 +29,19 @@ parameters:
      pt_BR: Imagem prompt, você pode verificar a documentação oficial do DallE 3
    llm_description: Image prompt of DallE 3, you should describe the image you want to generate as a list of words as possible as detailed
    form: llm
+  - name: seed_id
+    type: string
+    required: false
+    label:
+      en_US: Seed ID
+      zh_Hans: 种子ID
+      pt_BR: ID da semente
+    human_description:
+      en_US: Image generation seed ID to ensure consistency of series generated images
+      zh_Hans: 图像生成种子ID，确保系列生成图像的一致性
+      pt_BR: ID de semente de geração de imagem para garantir a consistência das imagens geradas em série
+    llm_description: If the user requests image consistency, extract the seed ID from the user's question or context.The seed id consists of an 8-bit string containing uppercase and lowercase letters and numbers
+    form: llm
  - name: size
    type: select
    required: true
@@ -1,3 +1,4 @@
+import random
 from base64 import b64decode
 from typing import Any, Union

@@ -9,10 +10,10 @@ from core.tools.tool.builtin_tool import BuiltinTool


 class DallE3Tool(BuiltinTool):
-    def _invoke(self, 
-                user_id: str, 
-               tool_parameters: dict[str, Any], 
-        ) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]:
+    def _invoke(self,
+                user_id: str,
+                tool_parameters: dict[str, Any],
+                ) -> Union[ToolInvokeMessage, list[ToolInvokeMessage]]:
        """
            invoke tools
        """
@@ -68,8 +69,13 @@ class DallE3Tool(BuiltinTool):
        result = []

        for image in response.data:
-            result.append(self.create_blob_message(blob=b64decode(image.b64_json), 
-                                                   meta={ 'mime_type': 'image/png' },
-                                                    save_as=self.VARIABLE_KEY.IMAGE.value))
-
+            result.append(self.create_blob_message(blob=b64decode(image.b64_json),
+                                                   meta={'mime_type': 'image/png'},
+                                                   save_as=self.VARIABLE_KEY.IMAGE.value))
        return result
+
+    @staticmethod
+    def _generate_random_id(length=8):
+        characters = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789'
+        random_id = ''.join(random.choices(characters, k=length))
+        return random_id
@@ -1,5 +1,5 @@
 from core.tools.errors import ToolProviderCredentialValidationError
-from core.tools.provider.builtin.duckduckgo.tools.duckduckgo_search import DuckDuckGoSearchTool
+from core.tools.provider.builtin.duckduckgo.tools.ddgo_search import DuckDuckGoSearchTool
 from core.tools.provider.builtin_tool_provider import BuiltinToolProviderController


@@ -0,0 +1,20 @@
+from typing import Any
+
+from duckduckgo_search import DDGS
+
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from core.tools.tool.builtin_tool import BuiltinTool
+
+
+class DuckDuckGoAITool(BuiltinTool):
+    """
+    Tool for performing a search using DuckDuckGo search engine.
+    """
+
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage:
+        query_dict = {
+            "keywords": tool_parameters.get('query'),
+            "model": tool_parameters.get('model'),
+        }
+        response = DDGS().chat(**query_dict)
+        return self.create_text_message(text=response)
@@ -0,0 +1,41 @@
+identity:
+  name: ddgo_ai
+  author: hjlarry
+  label:
+    en_US: DuckDuckGo AI Chat
+    zh_Hans: DuckDuckGo AI聊天
+description:
+  human:
+    en_US: Use the anonymous private chat provided by DuckDuckGo.
+    zh_Hans: 使用DuckDuckGo提供的匿名私密聊天。
+  llm: Use the anonymous private chat provided by DuckDuckGo.
+parameters:
+  - name: query
+    type: string
+    required: true
+    label:
+      en_US: Chat Content
+      zh_Hans: 聊天内容
+    human_description:
+      en_US: The chat content.
+      zh_Hans: 要聊天的内容。
+    llm_description: Key words for chat
+    form: llm
+  - name: model
+    type: select
+    required: true
+    options:
+      - value: gpt-3.5
+        label:
+          en_US: GPT-3.5
+      - value: claude-3-haiku
+        label:
+          en_US: Claude 3
+    default: gpt-3.5
+    label:
+      en_US: Choose Model
+      zh_Hans: 选择模型
+    human_description:
+      en_US: used to select the model for AI chat.
+      zh_Hans: 用于选择使用AI聊天的模型
+    form: form
@@ -0,0 +1,25 @@
+from typing import Any
+
+from duckduckgo_search import DDGS
+
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from core.tools.tool.builtin_tool import BuiltinTool
+
+
+class DuckDuckGoImageSearchTool(BuiltinTool):
+    """
+    Tool for performing an image search using DuckDuckGo search engine.
+    """
+
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> list[ToolInvokeMessage]:
+        query_dict = {
+            "keywords": tool_parameters.get('query'),
+            "timelimit": tool_parameters.get('timelimit'),
+            "size": tool_parameters.get('size'),
+            "max_results": tool_parameters.get('max_results'),
+        }
+        response = DDGS().images(**query_dict)
+        results = []
+        for res in response:
+            results.append(self.create_image_message(image=res.get("image")))
+        return results
@@ -0,0 +1,88 @@
+identity:
+  name: ddgo_img
+  author: hjlarry
+  label:
+    en_US: DuckDuckGo Image Search
+    zh_Hans: DuckDuckGo 图片搜索
+description:
+  human:
+    en_US: Perform image searches on DuckDuckGo and get results.
+    zh_Hans: 在 DuckDuckGo 上进行图片搜索并获取结果。
+  llm: Perform image searches on DuckDuckGo and get results.
+parameters:
+  - name: query
+    type: string
+    required: true
+    label:
+      en_US: Query string
+      zh_Hans: 查询语句
+    human_description:
+      en_US: The search query.
+      zh_Hans: 搜索查询语句。
+    llm_description: Key words for searching
+    form: llm
+  - name: max_results
+    type: number
+    required: true
+    default: 3
+    label:
+      en_US: Max results
+      zh_Hans: 最大结果数量
+    human_description:
+      en_US: The max results.
+      zh_Hans: 最大结果数量
+    form: form
+  - name: timelimit
+    type: select
+    required: false
+    options:
+      - value: Day
+        label:
+          en_US: current day
+          zh_Hans: 当天
+      - value: Week
+        label:
+          en_US: current week
+          zh_Hans: 本周
+      - value: Month
+        label:
+          en_US: current month
+          zh_Hans: 当月
+      - value: Year
+        label:
+          en_US: current year
+          zh_Hans: 今年
+    label:
+      en_US: Result time limit
+      zh_Hans: 结果时间限制
+    human_description:
+      en_US: Use when querying results within a specific time range only.
+      zh_Hans: 只查询一定时间范围内的结果时使用
+    form: form
+  - name: size
+    type: select
+    required: false
+    options:
+      - value: Small
+        label:
+          en_US: small
+          zh_Hans: 小
+      - value: Medium
+        label:
+          en_US: medium
+          zh_Hans: 中
+      - value: Large
+        label:
+          en_US: large
+          zh_Hans: 大
+      - value: Wallpaper
+        label:
+          en_US: xl
+          zh_Hans: 超大
+    label:
+      en_US: image size
+      zh_Hans: 图片大小
+    human_description:
+      en_US: The size of the image to be searched.
+      zh_Hans: 要搜索的图片的大小
+    form: form
@@ -0,0 +1,29 @@
+from typing import Any
+
+from duckduckgo_search import DDGS
+
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from core.tools.tool.builtin_tool import BuiltinTool
+
+
+class DuckDuckGoSearchTool(BuiltinTool):
+    """
+    Tool for performing a search using DuckDuckGo search engine.
+    """
+
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage:
+        query = tool_parameters.get('query', '')
+        result_type = tool_parameters.get('result_type', 'text')
+        max_results = tool_parameters.get('max_results', 10)
+        require_summary = tool_parameters.get('require_summary', False)
+        response = DDGS().text(query, max_results=max_results)
+
+        if result_type == 'link':
+            results = [f"[{res.get('title')}]({res.get('href')})" for res in response]
+            results = "\n".join(results)
+            return self.create_link_message(link=results)
+        results = [res.get("body") for res in response]
+        results = "\n".join(results)
+        if require_summary:
+            results = self.summary(user_id=user_id, content=results)
+        return self.create_text_message(text=results)
@@ -0,0 +1,65 @@
+identity:
+  name: ddgo_search
+  author: Yash Parmar
+  label:
+    en_US: DuckDuckGo Search
+    zh_Hans: DuckDuckGo 搜索
+description:
+  human:
+    en_US: Perform searches on DuckDuckGo and get results.
+    zh_Hans: 在 DuckDuckGo 上进行搜索并获取结果。
+  llm: Perform searches on DuckDuckGo and get results.
+parameters:
+  - name: query
+    type: string
+    required: true
+    label:
+      en_US: Query string
+      zh_Hans: 查询语句
+    human_description:
+      en_US: The search query.
+      zh_Hans: 搜索查询语句。
+    llm_description: Key words for searching
+    form: llm
+  - name: max_results
+    type: number
+    required: true
+    default: 5
+    label:
+      en_US: Max results
+      zh_Hans: 最大结果数量
+    human_description:
+      en_US: The max results.
+      zh_Hans: 最大结果数量
+    form: form
+  - name: result_type
+    type: select
+    required: true
+    options:
+      - value: text
+        label:
+          en_US: text
+          zh_Hans: 文本
+      - value: link
+        label:
+          en_US: link
+          zh_Hans: 链接
+    default: text
+    label:
+      en_US: Result type
+      zh_Hans: 结果类型
+    human_description:
+      en_US: used for selecting the result type, text or link
+      zh_Hans: 用于选择结果类型，使用文本还是链接进行展示
+    form: form
+  - name: require_summary
+    type: boolean
+    required: true
+    default: false
+    label:
+      en_US: Require Summary
+      zh_Hans: 是否总结
+    human_description:
+      en_US: Whether to pass the search results to llm for summarization.
+      zh_Hans: 是否需要将搜索结果传给大模型总结
+    form: form
@@ -0,0 +1,20 @@
+from typing import Any
+
+from duckduckgo_search import DDGS
+
+from core.tools.entities.tool_entities import ToolInvokeMessage
+from core.tools.tool.builtin_tool import BuiltinTool
+
+
+class DuckDuckGoTranslateTool(BuiltinTool):
+    """
+    Tool for performing a search using DuckDuckGo search engine.
+    """
+
+    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage:
+        query_dict = {
+            "keywords": tool_parameters.get('query'),
+            "to": tool_parameters.get('translate_to'),
+        }
+        response = DDGS().translate(**query_dict)[0].get('translated', 'Unable to translate!')
+        return self.create_text_message(text=response)
@@ -0,0 +1,51 @@
+identity:
+  name: ddgo_translate
+  author: hjlarry
+  label:
+    en_US: DuckDuckGo Translate
+    zh_Hans: DuckDuckGo 翻译
+description:
+  human:
+    en_US: Use DuckDuckGo's translation feature.
+    zh_Hans: 使用DuckDuckGo的翻译功能。
+  llm: Use DuckDuckGo's translation feature.
+parameters:
+  - name: query
+    type: string
+    required: true
+    label:
+      en_US: Translate Content
+      zh_Hans: 翻译内容
+    human_description:
+      en_US: The translate content.
+      zh_Hans: 要翻译的内容。
+    llm_description: Key words for translate
+    form: llm
+  - name: translate_to
+    type: select
+    required: true
+    options:
+      - value: en
+        label:
+          en_US: English
+          zh_Hans: 英语
+      - value: zh-Hans
+        label:
+          en_US: Simplified Chinese
+          zh_Hans: 简体中文
+      - value: zh-Hant
+        label:
+          en_US: Traditional Chinese
+          zh_Hans: 繁体中文
+      - value: ja
+        label:
+          en_US: Japanese
+          zh_Hans: 日语
+    default: en
+    label:
+      en_US: Choose Language
+      zh_Hans: 选择语言
+    human_description:
+      en_US: select the language to translate.
+      zh_Hans: 选择要翻译的语言
+    form: form
@@ -1,171 +0,0 @@
-from typing import Any, Optional
-
-from pydantic import BaseModel, Field
-
-from core.tools.entities.tool_entities import ToolInvokeMessage
-from core.tools.tool.builtin_tool import BuiltinTool
-
-
-class DuckDuckGoSearchAPIWrapper(BaseModel):
-    """Wrapper for DuckDuckGo Search API.
-
-    Free and does not require any setup.
-    """
-
-    region: Optional[str] = "wt-wt"
-    safesearch: str = "moderate"
-    time: Optional[str] = "y"
-    max_results: int = 5
-
-    def get_snippets(self, query: str) -> list[str]:
-        """Run query through DuckDuckGo and return concatenated results."""
-        from duckduckgo_search import DDGS
-
-        with DDGS() as ddgs:
-            results = ddgs.text(
-                query,
-                region=self.region,
-                safesearch=self.safesearch,
-                timelimit=self.time,
-            )
-            if results is None:
-                return ["No good DuckDuckGo Search Result was found"]
-            snippets = []
-            for i, res in enumerate(results, 1):
-                if res is not None:
-                    snippets.append(res["body"])
-                if len(snippets) == self.max_results:
-                    break
-        return snippets
-
-    def run(self, query: str) -> str:
-        snippets = self.get_snippets(query)
-        return " ".join(snippets)
-
-    def results(
-        self, query: str, num_results: int, backend: str = "api"
-    ) -> list[dict[str, str]]:
-        """Run query through DuckDuckGo and return metadata.
-
-        Args:
-            query: The query to search for.
-            num_results: The number of results to return.
-
-        Returns:
-            A list of dictionaries with the following keys:
-                snippet - The description of the result.
-                title - The title of the result.
-                link - The link to the result.
-        """
-        from duckduckgo_search import DDGS
-
-        with DDGS() as ddgs:
-            results = ddgs.text(
-                query,
-                region=self.region,
-                safesearch=self.safesearch,
-                timelimit=self.time,
-                backend=backend,
-            )
-            if results is None:
-                return [{"Result": "No good DuckDuckGo Search Result was found"}]
-
-            def to_metadata(result: dict) -> dict[str, str]:
-                if backend == "news":
-                    return {
-                        "date": result["date"],
-                        "title": result["title"],
-                        "snippet": result["body"],
-                        "source": result["source"],
-                        "link": result["url"],
-                    }
-                return {
-                    "snippet": result["body"],
-                    "title": result["title"],
-                    "link": result["href"],
-                }
-
-            formatted_results = []
-            for i, res in enumerate(results, 1):
-                if res is not None:
-                    formatted_results.append(to_metadata(res))
-                if len(formatted_results) == num_results:
-                    break
-        return formatted_results
-
-
-class DuckDuckGoSearchRun(BaseModel):
-    """Tool that queries the DuckDuckGo search API."""
-
-    name = "duckduckgo_search"
-    description = (
-        "A wrapper around DuckDuckGo Search. "
-        "Useful for when you need to answer questions about current events. "
-        "Input should be a search query."
-    )
-    api_wrapper: DuckDuckGoSearchAPIWrapper = Field(
-        default_factory=DuckDuckGoSearchAPIWrapper
-    )
-
-    def _run(
-        self,
-        query: str,
-    ) -> str:
-        """Use the tool."""
-        return self.api_wrapper.run(query)
-
-
-class DuckDuckGoSearchResults(BaseModel):
-    """Tool that queries the DuckDuckGo search API and gets back json."""
-
-    name = "DuckDuckGo Results JSON"
-    description = (
-        "A wrapper around Duck Duck Go Search. "
-        "Useful for when you need to answer questions about current events. "
-        "Input should be a search query. Output is a JSON array of the query results"
-    )
-    num_results: int = 4
-    api_wrapper: DuckDuckGoSearchAPIWrapper = Field(
-        default_factory=DuckDuckGoSearchAPIWrapper
-    )
-    backend: str = "api"
-
-    def _run(
-        self,
-        query: str,
-    ) -> str:
-        """Use the tool."""
-        res = self.api_wrapper.results(query, self.num_results, backend=self.backend)
-        res_strs = [", ".join([f"{k}: {v}" for k, v in d.items()]) for d in res]
-        return ", ".join([f"[{rs}]" for rs in res_strs])
-
-class DuckDuckGoInput(BaseModel):
-    query: str = Field(..., description="Search query.")
-
-class DuckDuckGoSearchTool(BuiltinTool):
-    """
-    Tool for performing a search using DuckDuckGo search engine.
-    """
-
-    def _invoke(self, user_id: str, tool_parameters: dict[str, Any]) -> ToolInvokeMessage | list[ToolInvokeMessage]:
-        """
-        Invoke the DuckDuckGo search tool.
-
-        Args:
-            user_id (str): The ID of the user invoking the tool.
-            tool_parameters (dict[str, Any]): The parameters for the tool invocation.
-
-        Returns:
-            ToolInvokeMessage | list[ToolInvokeMessage]: The result of the tool invocation.
-        """
-        query = tool_parameters.get('query', '')
-
-        if not query:
-            return self.create_text_message('Please input query')
-
-        tool = DuckDuckGoSearchRun(args_schema=DuckDuckGoInput)
-
-        result = tool._run(query)
-
-        return self.create_text_message(self.summary(user_id=user_id, content=result))
-    
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
jyong	909cc5a49f	redis iam check	2024-06-14 03:07:17 +08:00
Bowen LiangandGitHub	cdc08a434f	feat: support Chroma vector store (#5015 )	2024-06-13 18:02:18 +08:00
Kazuki TakamatsuandGitHub	3f18369ad2	Fix: google storage init with sa and download (#5054 )	2024-06-13 17:36:34 +08:00
Kazuki HasegawaandGitHub	db976a1f74	Upgrade boto3 library to support EKS Pod Identity. (#5064 )	2024-06-13 17:36:14 +08:00
kurokoboandGitHub	e61f5d029a	chore(docs): fix minor small typos (#5124 )	2024-06-13 17:36:01 +08:00
Charles ZhouandGitHub	eaca892c4e	fix: front end error when same tool is called twice at once (#5068 )	2024-06-13 17:16:59 +08:00
非法操作andGitHub	015c26d303	fix: style misalignment and inconsistency (#5149 )	2024-06-13 16:32:42 +08:00
sinoandGitHub	8210637bc5	feat: support jina-clip-v1 embedding model (#5146 )	2024-06-13 16:31:18 +08:00
呆萌闷油瓶andGitHub	790543131a	chore:add some new api version for azure openai (#5142 )	2024-06-13 16:30:47 +08:00
Ikko Eltociear AshimineandGitHub	a40f68cf94	chore: update qdrant_vector.py (#5128 )	2024-06-13 15:35:14 +08:00
yanghx GitHub -LAN-	adc948e87c	fix(api/core/model_runtime/model_providers/baichuan,localai): Parse ToolPromptMessage. #4943 (#5138 ) Co-authored-by: -LAN- <[email protected]>	2024-06-13 13:08:30 +08:00
smokyandGitHub	742b08e1d5	chore: update question classifier prompt (#5137 ) Signed-off-by: 0xff-dev <[email protected]>	2024-06-13 13:04:51 +08:00
orangeclkandGitHub	79e8489942	feat: support siliconflow (#5129 )	2024-06-13 12:59:41 +08:00
Charlie.Wei GitHubluowei <glpat-EjySCyNjWiLqAED-YmwM>crazywoola crazywoola	d6fa130cb5	remove dalle3 seed (#5136 ) Co-authored-by: luowei <glpat-EjySCyNjWiLqAED-YmwM> Co-authored-by: crazywoola <[email protected]> Co-authored-by: crazywoola <[email protected]>	2024-06-13 08:05:55 +08:00
takatostandGitHub	0c92f81efc	chore: sync pyproject.toml from requirements.txt (#5130 )	2024-06-13 00:06:05 +08:00
fishisnow GitHub huangyusong	11fd4a5dcc	Fix: fix load_yaml logging, Avoid setting the log level to warning (#5019 ) Co-authored-by: huangyusong <[email protected]>	2024-06-12 19:27:01 +08:00
Harry WangandGitHub	b399e8a359	fixed a typo and grammar error in sampled app (#5061 )	2024-06-12 18:02:22 +08:00
非法操作andGitHub	e04fc9b304	fix: select field not work when it is not required (#5101 )	2024-06-12 17:46:53 +08:00
xielong GitHub takatost Bowen Liang	ea69dc2a7e	feat: support hunyuan llm models (#5013 ) Co-authored-by: takatost <[email protected]> Co-authored-by: Bowen Liang <[email protected]>	2024-06-12 17:24:23 +08:00
PikaandGitHub	ecc7f130b4	fix(typo): misspelling (#5094 )	2024-06-12 17:01:21 +08:00
zxhlyhandGitHub	95443bd551	chore: workflow syncing modal (#5108 )	2024-06-12 16:35:19 +08:00
sinoandGitHub	0ce97e6315	feat: support doubao llm function calling (#5100 )	2024-06-12 15:43:50 +08:00
Bowen LiangandGitHub	25b0a97851	build: use Poetry as default build system for dependency installation in CI jobs (#5088 )	2024-06-12 14:43:03 +08:00
reroreroandGitHub	28997772a5	fix: remote_url doesn't work for gemini (#5090 )	2024-06-12 13:14:53 +08:00
Charlie.Wei GitHubluowei <glpat-EjySCyNjWiLqAED-YmwM>crazywoola crazywoola	b7c72f7a97	dalle3 add style consistency parameter (#5067 ) Co-authored-by: luowei <glpat-EjySCyNjWiLqAED-YmwM> Co-authored-by: crazywoola <[email protected]> Co-authored-by: crazywoola <[email protected]>	2024-06-12 12:59:03 +08:00
JoelandGitHub	9f7b38c068	fix: #4970 (#5093 )	2024-06-12 11:29:38 +08:00
非法操作andGitHub	3b36ba797f	feat: add duckduckgo img search, translate, ai chat (#5074 )	2024-06-12 10:04:10 +08:00
ugyujiandGitHub	4d2e6c3391	fix: Google HL Parameter for SearchApi (#5071 )	2024-06-12 08:30:01 +08:00
Nam VuandGitHub	3520d35f38	fix: autoHeightTextarea dimensions in Firefox (#4891 )	2024-06-12 08:24:58 +08:00
KVOJJJinandGitHub	5f104bab57	Fix: infinite loading not work when message is too short (#5075 )	2024-06-12 08:23:39 +08:00
orangeclkandGitHub	2050a8b8f0	feat: add glm4 new models and zhipu embedding-2 (#5089 )	2024-06-12 08:22:17 +08:00
takatostandGitHub	e3544c6ef7	fix: dependency package versions are not synchronized to requirements.txt (#5084 )	2024-06-11 22:21:18 +08:00
ra2230andGitHub	472b976946	Arabic README.md (#5078 )	2024-06-11 18:43:54 +08:00
Matri GitHub Bowen Liang	f62f71a81a	build: initial support for poetry build tool (#4513 ) Co-authored-by: Bowen Liang <[email protected]>	2024-06-11 13:11:28 +08:00
Takuya OnoandGitHub	f426e1b3bd	🔧 Fix(docker/volumes/ssrf_proxy/squid.conf): The squid process on ssrf_proxy docker service crashes at startup (#5050 )	2024-06-11 12:32:05 +08:00
sinoandGitHub	5f870ac950	chore: update maas model provider description (#5056 )	2024-06-11 11:22:22 +08:00
Pascal MandGitHub	415816cf35	feat: add dataset delete endpoint (#5048 )	2024-06-11 11:21:38 +08:00
Pascal MandGitHub	9103112555	fix: wrong link to web app repo in chatflow mode (#5062 )	2024-06-11 11:20:52 +08:00
takatostandGitHub	5986841e27	fix: issue where an error occurs when invoking TTS without selecting a voice (#5046 )	2024-06-09 20:28:24 +08:00
Jaxon LeyandGitHub	2573b138bf	fix: update presence_penalty configuration for wenxin AI ernie-4.0-8k and ernie-3.5-8k models (#5039 )	2024-06-09 14:44:11 +08:00
Takuya OnoandGitHub	308ce66af5	🔧 fix docker-compose ssrf_proxy service WARNING: You should probably remove '::/0' from the ACL named 'all' (#5005 )	2024-06-09 14:39:52 +08:00
Bowen LiangandGitHub	bdad993901	improve: generalize vector factory classes and vector type (#5033 )	2024-06-08 22:29:24 +08:00
zxhlyhandGitHub	3b62ab564a	feat: feature modal style (#5032 )	2024-06-08 07:32:34 +08:00
PikaandGitHub	d319d9fc5e	fix(style): some style issues (#5029 )	2024-06-07 20:59:39 +08:00
doufaandGitHub	ea5c8a72e2	Fix language setting not success (#5023 )	2024-06-07 20:02:08 +08:00
JyongandGitHub	3b60c28b3a	deal the external image when extract docx image (#5024 )	2024-06-07 20:00:39 +08:00
KVOJJJinandGitHub	ea0219a5d5	Fix: z-index in header (#5017 )	2024-06-07 16:01:33 +08:00
JyongandGitHub	481e7bc6b9	Fix/azure blob new version (#5004 )	2024-06-06 23:36:13 +08:00
Charles ZhouandGitHub	1ccba85c91	fix: modal z-index and cleanup (#4978 )	2024-06-06 22:28:13 +08:00
doufaandGitHub	2539e56514	fix: some base models cannot be selected in Azure OpenAI Service setting page (#4985 )	2024-06-06 22:27:57 +08:00
takatostandGitHub	3929d289e0	feat: set default memory messages limit to infinite (#5002 )	2024-06-06 17:39:44 +08:00
YeuolyandGitHub	52585aea74	fix: typo in sd3 (#5000 )	2024-06-06 17:08:49 +08:00
YeuolyandGitHub	73dee84cab	fix: add handling for non-string type in variable template parser (#4996 )	2024-06-06 16:38:13 +08:00
JoelandGitHub	efecdccf35	feat: support login by given mail (#4991 )	2024-06-06 15:01:58 +08:00
JoelandGitHub	da5f2e168a	fix: llm selector position is incorrect in not workflow app (#4982 )	2024-06-06 10:47:36 +08:00
JoeandGitHub	5cdb95be1f	fix: gemini timeout error (#4955 )	2024-06-06 10:19:03 +08:00
Bowen LiangandGitHub	7fa735a43b	chore: rename vdb tests for PGVector and PGvectoRS (#4973 )	2024-06-06 07:22:49 +08:00
takatostandGitHub	3579fd1b09	feat: add create tenant command (#4974 )	2024-06-06 00:42:00 +08:00
JyongandGitHub	237b8fe3d9	add meta.doc_id index for tidb (#4963 )	2024-06-05 20:45:43 +08:00
JyongandGitHub	02e4de5166	fix some tidb bugs (#4960 )	2024-06-05 19:14:18 +08:00
Mark SunandGitHub	64c8093c1e	Typo in Knowledge settings (#4958 )	2024-06-05 18:31:24 +08:00
WeaxsandGitHub	0797f9bc05	feat: support tidb vector (#4588 )	2024-06-05 18:19:53 +08:00
非法操作 GitHub Jyong	602c4e51ec	fix: duckduckgo search does not work (#4949 ) Co-authored-by: Jyong <[email protected]>	2024-06-05 17:33:58 +08:00
YCandGitHub	9f8ca75a81	fixing a bug of handling header row when parsing xls file, and tune xls/xlsx parsing result to be more structured (#3600 )	2024-06-05 15:28:43 +08:00
YeuolyandGitHub	80a87f36ea	fix: missing iterator in task pipeline (#4948 )	2024-06-05 15:10:20 +08:00
Charles ZhouandGitHub	63addc9258	fix: missing dataset patch parameters in settings modal (#4901 )	2024-06-05 14:21:59 +08:00
Bowen LiangandGitHub	f32b440c4a	chore: fix indention violations by applying E111 to E117 ruff rules (#4925 )	2024-06-05 14:05:15 +08:00
crazywoolaandGitHub	6b6afb7708	fix: import error in web/app/components/header/account-setting/model-provider-page/declarations.ts (#4944 )	2024-06-05 14:01:12 +08:00
zxhlyhandGitHub	a4041cb40b	fix: end node limit in next step (#4945 )	2024-06-05 14:00:47 +08:00
JasonVVandGitHub	7749b71fff	Optimize knowledge retrieval performance by batching dataset quries. (#4917 )	2024-06-05 13:30:32 +08:00
JoelandGitHub	3006124e6d	feat: pricing page add llm load balancing info (#4942 )	2024-06-05 11:31:44 +08:00
Harry WangandGitHub	3d276f4a7f	change "Import from text file" to "Import from file" (#4935 )	2024-06-05 09:29:29 +08:00
				`@@ -0,0 +1 @@`
				<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" fill="none" version="1.1" width="128" height="128" viewBox="0 0 128 128"><g><g style="opacity:0;"><rect x="0" y="0" width="128" height="128" rx="0" fill="#FFFFFF" fill-opacity="1"/></g><g><path d="M100.74,12L93.2335,12C69.21260000000001,12,55.3672,27.3468,55.3672,50.8672L55.3672,54.8988C52.6011,54.1056,49.7377,53.7031,46.8601,53.7031C29.816499999999998,53.7031,16,67.5196,16,84.5632C16,101.6069,29.816499999999998,115.423,46.8601,115.423C63.9037,115.423,77.72030000000001,101.6069,77.72030000000001,84.5632C77.72030000000001,82.4902,77.51140000000001,80.4223,77.0967,78.3911L77.2197,78.3911L100.74,78.3911C106.9654,78.3681,112,73.3151,112,67.08959999999999C112,60.8642,106.9654,55.8111,100.74,55.7882L100.7362,55.7882L100.6985,55.7879L100.6606,55.7882L77.2197,55.7882L77.2195,49.8663C77.2195,40.8584,83.7252,34.352900000000005,93.2335,34.352900000000005L100.5653,34.352900000000005L100.5733,34.352900000000005L100.5812,34.352900000000005L100.74,34.352900000000005L100.74,34.352900000000005C106.8469,34.2605,111.7497,29.284,111.7497,23.1764C111.7497,17.06889,106.8469,12.0923454,100.74,12L100.74,12ZM56.0347,84.5632C56.0347,79.4962,51.9271,75.3885,46.8601,75.3885C41.793099999999995,75.3885,37.6854,79.4962,37.6854,84.5632C37.6854,89.6303,41.793099999999995,93.7378,46.8601,93.7378C51.9271,93.7378,56.0347,89.6303,56.0347,84.5632Z" fill-rule="evenodd" fill="#8358F6" fill-opacity="1"/></g></g></svg>