Feat: Create a conversation #3221 (#9269)

### What problem does this PR solve?

Feat: Create a conversation #3221

### Type of change


- [x] New Feature (non-breaking change which adds functionality)

.gitignore

@@ -193,3 +193,5 @@ dist
 # SvelteKit build / generate output
 .svelte-kit
 
+# Default backup dir
+backup

.trivyignore

@@ -0,0 +1,15 @@
+**/*.md
+**/*.min.js
+**/*.min.css
+**/*.svg
+**/*.png
+**/*.jpg
+**/*.jpeg
+**/*.gif
+**/*.woff
+**/*.woff2
+**/*.map
+**/*.webp
+**/*.ico
+**/*.ttf
+**/*.eot

README.md

@@ -87,7 +87,8 @@ Try our demo at [https://demo.ragflow.io](https://demo.ragflow.io).
 
 ## 🔥 Latest Updates
 
-- 2025-08-01 Supports agentic workflow.
+- 2025-08-04 Supports new models, including Kimi K2 and Grok 4.
+- 2025-08-01 Supports agentic workflow and MCP.
 - 2025-05-23 Adds a Python/JavaScript code executor component to Agent.
 - 2025-05-05 Supports cross-language query.
 - 2025-03-19 Supports using a multi-modal model to make sense of images within PDF or DOCX files.

README_id.md

@@ -80,7 +80,8 @@ Coba demo kami di [https://demo.ragflow.io](https://demo.ragflow.io).
 
 ## 🔥 Pembaruan Terbaru
 
-- 2025-08-01 Mendukung Alur Kerja agen.
+- 2025-08-04 Mendukung model baru, termasuk Kimi K2 dan Grok 4.
+- 2025-08-01 Mendukung alur kerja agen dan MCP.
 - 2025-05-23 Menambahkan komponen pelaksana kode Python/JS ke Agen.
 - 2025-05-05 Mendukung kueri lintas bahasa.
 - 2025-03-19 Mendukung penggunaan model multi-modal untuk memahami gambar di dalam file PDF atau DOCX.

README_ja.md

@@ -60,7 +60,8 @@
 
 ## 🔥 最新情報
 
-- 2025-08-01 エージェントワークフローをサポートします。
+- 2025-08-04 新モデル、キミK2およびGrok 4をサポート。
+- 2025-08-01 エージェントワークフローとMCPをサポート。
 - 2025-05-23 エージェントに Python/JS コードエグゼキュータコンポーネントを追加しました。
 - 2025-05-05 言語間クエリをサポートしました。
 - 2025-03-19 PDFまたはDOCXファイル内の画像を理解するために、多モーダルモデルを使用することをサポートします。

README_ko.md

@@ -60,7 +60,8 @@
 
 ## 🔥 업데이트
 
-- 2025-08-01 에이전트 워크플로를 지원합니다.
+- 2025-08-04 새로운 모델인 Kimi K2와 Grok 4를 포함하여 지원합니다.
+- 2025-08-01 에이전트 워크플로우와 MCP를 지원합니다.
 - 2025-05-23 Agent에 Python/JS 코드 실행기 구성 요소를 추가합니다.
 - 2025-05-05 언어 간 쿼리를 지원합니다.
 - 2025-03-19 PDF 또는 DOCX 파일 내의 이미지를 이해하기 위해 다중 모드 모델을 사용하는 것을 지원합니다.

README_pt_br.md

@@ -80,7 +80,8 @@ Experimente nossa demo em [https://demo.ragflow.io](https://demo.ragflow.io).
 
 ## 🔥 Últimas Atualizações
 
-- 01-08-2025 Suporta o fluxo de trabalho agêntico.
+- 04-08-2025 Suporta novos modelos, incluindo Kimi K2 e Grok 4.
+- 01-08-2025 Suporta fluxo de trabalho agente e MCP.
 - 23-05-2025 Adicione o componente executor de código Python/JS ao Agente.
 - 05-05-2025 Suporte a consultas entre idiomas.
 - 19-03-2025 Suporta o uso de um modelo multi-modal para entender imagens dentro de arquivos PDF ou DOCX.

README_tzh.md

@@ -83,7 +83,8 @@
 
 ## 🔥 近期更新
 
-- 2025-08-01 支援 agentic workflow
+- 2025-08-04 支援 Kimi K2 和 Grok 4 等模型.
+- 2025-08-01 支援 agentic workflow 和 MCP
 - 2025-05-23 為 Agent 新增 Python/JS 程式碼執行器元件。
 - 2025-05-05 支援跨語言查詢。
 - 2025-03-19 PDF和DOCX中的圖支持用多模態大模型去解析得到描述.

README_zh.md

@@ -83,7 +83,8 @@
 
 ## 🔥 近期更新
 
-- 2025-08-01 支持 agentic workflow。
+- 2025-08-04 新增对 Kimi K2 和 Grok 4 等模型的支持.
+- 2025-08-01 支持 agentic workflow 和 MCP。
 - 2025-05-23 Agent 新增 Python/JS 代码执行器组件。
 - 2025-05-05 支持跨语言查询。
 - 2025-03-19 PDF 和 DOCX 中的图支持用多模态大模型去解析得到描述.

agent/templates/customer_support.json

@@ -89,11 +89,11 @@
                             "presence_penalty": 0.4,
                             "prompts": [
                                 {
-                                    "content": "{sys.query}",
+                                    "content": "The user query is {sys.query}\n\nThe relevant document are {Retrieval:ShyPumasJoke@formalized_content}",
                                     "role": "user"
                                 }
                             ],
-                            "sys_prompt": "You are a highly professional product information advisor.  \n\nYour only mission is to provide accurate, factual, and structured answers to all product-related queries.\n\nAbsolutely no assumptions, guesses, or fabricated content are allowed. \n\n**Key Principles:**\n\n1. **Strict Database Reliance:**  \n\n   - Every answer must be based solely on the verified product information stored in the database accessed through the Retrieval tool.  \n\n   - You are NOT allowed to invent, speculate, or infer details beyond what is retrieved.  \n\n   - If you cannot find relevant data, respond with: *\"I cannot find this information in our official product database. Please check back later or provide more details for further search.\"*\n\n2. **Information Accuracy and Structure:**  \n\n   - Provide information in a clear, concise, and professional way.  \n\n   - Use bullet points or numbered lists if there are multiple key points (e.g., features, price, warranty, technical specifications).  \n\n   - Always specify the version or model number when applicable to avoid confusion.\n\n3. **Tone and Style:**  \n\n   - Maintain a polite, professional, and helpful tone at all times.  \n\n   - Avoid marketing exaggeration or promotional language; stay strictly factual.  \n\n   - Do not express personal opinions; only cite official product data.\n\n4.  **User Guidance:**  \n\n   - If the user\u2019s query is unclear or too broad, politely request clarification or guide them to provide more specific product details (e.g., product name, model, version).  \n\n   - Example: *\"Could you please specify the product model or category so I can retrieve the most relevant information for you?\"*\n\n5. **Response Length and Formatting:**  \n\n   - Keep each answer within 100\u2013150 words for general queries.  \n\n   - For complex or multi-step explanations, you may extend to 200\u2013250 words, but always remain clear and well-structured.\n\n6. **Critical Reminder:**  \n\nYour authority and reliability depend entirely on database-driven responses. Any fabricated, speculative, or unverified content will be considered a critical failure of your role.\n\nAlways begin processing a query by accessing the Retrieval tool, confirming the data source, and then structuring your response according to the above principles.\n\n",
+                            "sys_prompt": "You are a highly professional product information advisor.  \n\nYour only mission is to provide accurate, factual, and structured answers to all product-related queries.\n\nAbsolutely no assumptions, guesses, or fabricated content are allowed. \n\n**Key Principles:**\n\n1. **Strict Database Reliance:**  \n\n   - Every answer must be based solely on the verified product information stored in the relevant documen.\n\n   - You are NOT allowed to invent, speculate, or infer details beyond what is retrieved.  \n\n   - If you cannot find relevant data, respond with: *\"I cannot find this information in our official product database. Please check back later or provide more details for further search.\"*\n\n2. **Information Accuracy and Structure:**  \n\n   - Provide information in a clear, concise, and professional way.  \n\n   - Use bullet points or numbered lists if there are multiple key points (e.g., features, price, warranty, technical specifications).  \n\n   - Always specify the version or model number when applicable to avoid confusion.\n\n3. **Tone and Style:**  \n\n   - Maintain a polite, professional, and helpful tone at all times.  \n\n   - Avoid marketing exaggeration or promotional language; stay strictly factual.  \n\n   - Do not express personal opinions; only cite official product data.\n\n4.  **User Guidance:**  \n\n   - If the user\u2019s query is unclear or too broad, politely request clarification or guide them to provide more specific product details (e.g., product name, model, version).  \n\n   - Example: *\"Could you please specify the product model or category so I can retrieve the most relevant information for you?\"*\n\n5. **Response Length and Formatting:**  \n\n   - Keep each answer within 100\u2013150 words for general queries.  \n\n   - For complex or multi-step explanations, you may extend to 200\u2013250 words, but always remain clear and well-structured.\n\n6. **Critical Reminder:**  \n\nYour authority and reliability depend entirely on the relevant document responses. Any fabricated, speculative, or unverified content will be considered a critical failure of your role.\n\n\n",
                             "temperature": 0.1,
                             "temperatureEnabled": true,
                             "tools": [],
@@ -699,7 +699,7 @@
                             "width": 200
                         },
                         "position": {
-                            "x": 644.5771854408022,
+                            "x": 645.6873721057459,
                             "y": 516.6923702571407
                         },
                         "selected": false,
@@ -735,11 +735,11 @@
                                 "presence_penalty": 0.4,
                                 "prompts": [
                                     {
-                                        "content": "{sys.query}",
+                                        "content": "The user query is {sys.query}\n\nThe relevant document are {Retrieval:ShyPumasJoke@formalized_content}",
                                         "role": "user"
                                     }
                                 ],
-                                "sys_prompt": "You are a highly professional product information advisor.  \n\nYour only mission is to provide accurate, factual, and structured answers to all product-related queries.\n\nAbsolutely no assumptions, guesses, or fabricated content are allowed. \n\n**Key Principles:**\n\n1. **Strict Database Reliance:**  \n\n   - Every answer must be based solely on the verified product information stored in the database accessed through the Retrieval tool.  \n\n   - You are NOT allowed to invent, speculate, or infer details beyond what is retrieved.  \n\n   - If you cannot find relevant data, respond with: *\"I cannot find this information in our official product database. Please check back later or provide more details for further search.\"*\n\n2. **Information Accuracy and Structure:**  \n\n   - Provide information in a clear, concise, and professional way.  \n\n   - Use bullet points or numbered lists if there are multiple key points (e.g., features, price, warranty, technical specifications).  \n\n   - Always specify the version or model number when applicable to avoid confusion.\n\n3. **Tone and Style:**  \n\n   - Maintain a polite, professional, and helpful tone at all times.  \n\n   - Avoid marketing exaggeration or promotional language; stay strictly factual.  \n\n   - Do not express personal opinions; only cite official product data.\n\n4.  **User Guidance:**  \n\n   - If the user\u2019s query is unclear or too broad, politely request clarification or guide them to provide more specific product details (e.g., product name, model, version).  \n\n   - Example: *\"Could you please specify the product model or category so I can retrieve the most relevant information for you?\"*\n\n5. **Response Length and Formatting:**  \n\n   - Keep each answer within 100\u2013150 words for general queries.  \n\n   - For complex or multi-step explanations, you may extend to 200\u2013250 words, but always remain clear and well-structured.\n\n6. **Critical Reminder:**  \n\nYour authority and reliability depend entirely on database-driven responses. Any fabricated, speculative, or unverified content will be considered a critical failure of your role.\n\nAlways begin processing a query by accessing the Retrieval tool, confirming the data source, and then structuring your response according to the above principles.\n\n",
+                                "sys_prompt": "You are a highly professional product information advisor.  \n\nYour only mission is to provide accurate, factual, and structured answers to all product-related queries.\n\nAbsolutely no assumptions, guesses, or fabricated content are allowed. \n\n**Key Principles:**\n\n1. **Strict Database Reliance:**  \n\n   - Every answer must be based solely on the verified product information stored in the relevant documen.\n\n   - You are NOT allowed to invent, speculate, or infer details beyond what is retrieved.  \n\n   - If you cannot find relevant data, respond with: *\"I cannot find this information in our official product database. Please check back later or provide more details for further search.\"*\n\n2. **Information Accuracy and Structure:**  \n\n   - Provide information in a clear, concise, and professional way.  \n\n   - Use bullet points or numbered lists if there are multiple key points (e.g., features, price, warranty, technical specifications).  \n\n   - Always specify the version or model number when applicable to avoid confusion.\n\n3. **Tone and Style:**  \n\n   - Maintain a polite, professional, and helpful tone at all times.  \n\n   - Avoid marketing exaggeration or promotional language; stay strictly factual.  \n\n   - Do not express personal opinions; only cite official product data.\n\n4.  **User Guidance:**  \n\n   - If the user\u2019s query is unclear or too broad, politely request clarification or guide them to provide more specific product details (e.g., product name, model, version).  \n\n   - Example: *\"Could you please specify the product model or category so I can retrieve the most relevant information for you?\"*\n\n5. **Response Length and Formatting:**  \n\n   - Keep each answer within 100\u2013150 words for general queries.  \n\n   - For complex or multi-step explanations, you may extend to 200\u2013250 words, but always remain clear and well-structured.\n\n6. **Critical Reminder:**  \n\nYour authority and reliability depend entirely on the relevant document responses. Any fabricated, speculative, or unverified content will be considered a critical failure of your role.\n\n\n",
                                 "temperature": 0.1,
                                 "temperatureEnabled": true,
                                 "tools": [],

agent/templates/cv_evaluation.json

@@ -170,7 +170,7 @@
                             "presence_penalty": 0.5,
                             "prompts": [
                                 {
-                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}",
+                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Outline agent output is {Agent:BetterSitesSend@content}",
                                     "role": "user"
                                 }
                             ],
@@ -250,7 +250,7 @@
                             "presence_penalty": 0.5,
                             "prompts": [
                                 {
-                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
+                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Outline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
                                     "role": "user"
                                 }
                             ],
@@ -602,7 +602,7 @@
                                 "presence_penalty": 0.5,
                                 "prompts": [
                                     {
-                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}",
+                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Outline agent output is {Agent:BetterSitesSend@content}",
                                         "role": "user"
                                     }
                                 ],
@@ -715,7 +715,7 @@
                                 "presence_penalty": 0.5,
                                 "prompts": [
                                     {
-                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
+                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Outline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
                                         "role": "user"
                                     }
                                 ],

agent/templates/market_generate_seo_blog.json

@@ -169,7 +169,7 @@
                             "presence_penalty": 0.5,
                             "prompts": [
                                 {
-                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}",
+                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Outline agent output is {Agent:BetterSitesSend@content}",
                                     "role": "user"
                                 }
                             ],
@@ -249,7 +249,7 @@
                             "presence_penalty": 0.5,
                             "prompts": [
                                 {
-                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
+                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Outline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
                                     "role": "user"
                                 }
                             ],
@@ -601,7 +601,7 @@
                                 "presence_penalty": 0.5,
                                 "prompts": [
                                     {
-                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}",
+                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Outline agent output is {Agent:BetterSitesSend@content}",
                                         "role": "user"
                                     }
                                 ],
@@ -714,7 +714,7 @@
                                 "presence_penalty": 0.5,
                                 "prompts": [
                                     {
-                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
+                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Outline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
                                         "role": "user"
                                     }
                                 ],
@@ -912,4 +912,4 @@
             "retrieval": []
         },
     "avatar": "data:image/jpeg;base64,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"
-}
+}

agent/templates/seo_blog.json

@@ -169,7 +169,7 @@
                             "presence_penalty": 0.5,
                             "prompts": [
                                 {
-                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}",
+                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Outline agent output is {Agent:BetterSitesSend@content}",
                                     "role": "user"
                                 }
                             ],
@@ -249,7 +249,7 @@
                             "presence_penalty": 0.5,
                             "prompts": [
                                 {
-                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
+                                    "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Outline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
                                     "role": "user"
                                 }
                             ],
@@ -601,7 +601,7 @@
                                 "presence_penalty": 0.5,
                                 "prompts": [
                                     {
-                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}",
+                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\n\n\nThe Outline agent output is {Agent:BetterSitesSend@content}",
                                         "role": "user"
                                     }
                                 ],
@@ -714,7 +714,7 @@
                                 "presence_penalty": 0.5,
                                 "prompts": [
                                     {
-                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Ouline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
+                                        "content": "The parse and keyword agent output is {Agent:ClearRabbitsScream@content}\n\nThe Outline agent output is {Agent:BetterSitesSend@content}\n\nThe Body agent output is {Agent:EagerNailsRemain@content}",
                                         "role": "user"
                                     }
                                 ],
@@ -912,4 +912,4 @@
             "retrieval": []
         },
     "avatar": "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/4gHYSUNDX1BST0ZJTEUAAQEAAAHIAAAAAAQwAABtbnRyUkdCIFhZWiAH4AABAAEAAAAAAABhY3NwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAQAA9tYAAQAAAADTLQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAlkZXNjAAAA8AAAACRyWFlaAAABFAAAABRnWFlaAAABKAAAABRiWFlaAAABPAAAABR3dHB0AAABUAAAABRyVFJDAAABZAAAAChnVFJDAAABZAAAAChiVFJDAAABZAAAAChjcHJ0AAABjAAAADxtbHVjAAAAAAAAAAEAAAAMZW5VUwAAAAgAAAAcAHMAUgBHAEJYWVogAAAAAAAAb6IAADj1AAADkFhZWiAAAAAAAABimQAAt4UAABjaWFlaIAAAAAAAACSgAAAPhAAAts9YWVogAAAAAAAA9tYAAQAAAADTLXBhcmEAAAAAAAQAAAACZmYAAPKnAAANWQAAE9AAAApbAAAAAAAAAABtbHVjAAAAAAAAAAEAAAAMZW5VUwAAACAAAAAcAEcAbwBvAGcAbABlACAASQBuAGMALgAgADIAMAAxADb/2wBDAAEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQH/2wBDAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQEBAQH/wAARCAAwADADASIAAhEBAxEB/8QAGQAAAwEBAQAAAAAAAAAAAAAABgkKBwUI/8QAMBAAAAYCAQIEBQQCAwAAAAAAAQIDBAUGBxEhCAkAEjFBFFFhcaETFiKRFyOx8PH/xAAaAQACAwEBAAAAAAAAAAAAAAACAwABBgQF/8QALBEAAgIBAgUCBAcAAAAAAAAAAQIDBBEFEgATITFRIkEGIzJhFBUWgaGx8P/aAAwDAQACEQMRAD8AfF2hez9089t7pvxgQMa1Gb6qZ6oQE9m/NEvCIStyPfJSOF/M1epzMugo/qtMqbiRc1mJjoJKCLMNIxKcsLJedfO1Ct9cI63x9fx6CA/19t+oh4LFA5HfuAgP/A8eOIsnsTBrkBHXA7+v53+Q+ficTgJft9gIgA+/P9/1r342O/YA8A8k3/if+IbAN7+2/f8AAiI6H19PGoPyESTMZQPKUAHkQEN+3r9dh78/YPGUTk2wb/qAZZIugH1OHH5DjkdfbnWw2DsOxPj+xjrnx2H39unBopJGBn9s+PHv1HXjPJtH+J+B40O9a16h/wB/92j/ALrPa/wR104UyAobHlXhuo2HrEtK4qy3CwjKOuJLRHJLSkXWrFKs/gVrJVrE8TUiH8bPrP20UEu8m4hNpMJJuTOfnbUw/kUqyZgMHGjAO9+mtDsQ53sdcB6eMhnpEjhNQxRKICAgHy5+/roOdjr7c+J6O4x07dx484/n7nzw1gexBGfIPkZ/3t39uGpqc6+fP5/Ht8vGFZCzJjWpWuBxvO2yPjrtclUUK7BqmUI4fuASeyhG5FzFI0Bw4aQ0iZNoDgzvRW4qtyFkI4XmwyEk2YNnDp0sVBu3IUyy5iqH8gqKERSIRNIii67hddRJs1at01Xbx2sgzZoLu10UFJR+4V1A5cxF3FqNcLvjwcno43uuLrOxZYjujaClcb4QQfxEizpFiQyM9olcueRnjC2ZMt9iY06zL0qytrMSqSOVGsfHMaGhZ3l4lSRI2MqE74zJvRTveNFWWIh3RWw+XCAM5icKQLrCH57T17FhErSlRXnWvyZXKQwWJ3eraD14p5YuZCFgacskK2oGkVuKO5GYTHzf7DaD12cBD3DgPOIDrWw9PnrXPgDkpVsUDGMG+DD6E9gHXIjrYjwUPQTCXYgHPhIV974+F6E1hpC14Yzmzj56YaQEeZhXsayD1zLPW7pygxaMf81Nzu1iJsnIuDIKnaJAkPldqrHaoORZ73tMVEbFdSXT9nVgRQgnBq6j8e/HCIEATpAnH5KlmRVkFRFJwks/bqImSXJ5VFyA3N6Ikh3bCW3YHp5cowOmCfTgA+xJCnrjtwHKcLvJj2ZGcTRFj19kEhckdzgEjKnABGSSzdc1Fe5byXXGNjKdvRcw5NxvLidNZFFCxUa62KrzMaChw8hhYScFJtROAgmuLByq1MsgkZYPaVVuDe0wraRaqAdJwgRQo+YR8xTlAQNx6b49w41vXiJpCalLh1jZhyrTqRM4+jstdRmYryNkydLQRWg1LNGcWd5jIFFvCythlIySa0mNu74sKRQtaWsTmupqPItw0lE52ufpyYzrSkx6cw5bLmBEpkTsz+dt8P5QFuCRtAIkBH9MuwKHICIaDQhnojMs9mKaeGcrMxXlQtAYkdVljimRrE5MqI4zL8oSqQ6wxjodBqK05qdK3Vo3aCSVkBW7bjuC1NFJJBPaqyx6fp6pWkliYLXK2XrukkRu2CCVoSWMgsdMyySKwoLFcIGWSTUMg4IBgTcICoBhRcplMcpFkhIqQp1ClMBTmA0Zfe1zpjvHfXff65bZlzXpB3jjGTgiirmPjAfs16PHqHeQ75Wbj3xxZpOEkV3LRJJSPdomUBZISJLncV2k+8D07dxXp7xsYuTapA9UkJUYWIzNhadnWEZeCXGLQQiJi1ViHfhHL2unWh+mlORsrW0JFpEFnGVfm1mU4kq0FY3eD6corJncv6dr5NLSMNXVaTUksjTiMnaq8uFfSVuDyiJ1iZpy0LOJtpa3YfkcQ5fdozyxI2m5qqcrHN61YYmHsh6v3o9ParYmYJEtlhIx6+gUbjgD23M6oqg92YL0JyF6Bps+qDValVA9h9Lj5SZI3SHXdEQlj1wiQtLLIe6pGzjO3BlBkK1hxpblLVH5wdW0BcFKf/JwRtjsot2z8omaSdxbzzk1iEjsE0AM9rrRZNRIrVyo7dGO6E+oh8axLlJ5H5VaJKx7ePRGFbW6vUeFfHQIWPTI9Tm7HHfuhqY7E6C7JFqUzM6iZXIoncNxX7+bIVdJnTT48x3OQU1krIDW3UeixVhyISzYz6cadY5Xph6TseRNTRsTElzzBn9Vlly0TAERsdgnMYyLROjyFbg5R4ZlsGaMT4yNi2Zlq1GwjZB3jq0PsaJfA3t0jL0W0Y9xf1V41lpWckXMLaZiwxuKYPqc6LlHdkeRF+Qxswx5ASDqBVrsL+2A/N6SiCbYymV2BywJiMZj3GRRMTnL+lVyHCll3R7Szv0vqXMtQ74T+HijljIScLaEpkKCB3rqMBIi0jPs5JeOKTZMZEi5VVnouzy0k3jXjWSMlY6UcVGDxlKMVDqx91SILWSi3D2KdgYy3kP8E9X/AE1SnRXBNdNRMlefT6g7aY6giK+cPLGNg0bY68rcnpsNh9PqIBve/EcPQ3WIq2dR93xpSgk5SAZ9R6MLAOZFUkpLSUDXp6/KPpGUkmTdswlnKnwbl5ITMdGwcXJi7LKsqzUmT5tWYmkXuF9wjBvb76b7dHheazJ9RElUJOCxViuMlUJC0Gtz6PKyjLBY4qMWUe12r1xZ6lOyT6XPEBKN2CkTDOlZd02TBdTMt7Upx2knrkdCv1UKjDKn1A7XBYH6SCOOrWn5Oi/DtRiu+GleRthDL8rXdVjZlcfWrSIxVlGGGCOnH//Z"
-}
+}

agentic_reasoning/prompts.py

@@ -20,94 +20,128 @@ BEGIN_SEARCH_RESULT = "<|begin_search_result|>"
 END_SEARCH_RESULT = "<|end_search_result|>"
 MAX_SEARCH_LIMIT = 6
 
-REASON_PROMPT = (
-        "You are a reasoning assistant with the ability to perform dataset searches to help "
-        "you answer the user's question accurately. You have special tools:\n\n"
-        f"- To perform a search: write {BEGIN_SEARCH_QUERY} your query here {END_SEARCH_QUERY}.\n"
-        f"Then, the system will search and analyze relevant content, then provide you with helpful information in the format {BEGIN_SEARCH_RESULT} ...search results... {END_SEARCH_RESULT}.\n\n"
-        f"You can repeat the search process multiple times if necessary. The maximum number of search attempts is limited to {MAX_SEARCH_LIMIT}.\n\n"
-        "Once you have all the information you need, continue your reasoning.\n\n"
-        "-- Example 1 --\n" ########################################
-        "Question: \"Are both the directors of Jaws and Casino Royale from the same country?\"\n"
-        "Assistant:\n"
-        f"    {BEGIN_SEARCH_QUERY}Who is the director of Jaws?{END_SEARCH_QUERY}\n\n"
-        "User:\n"
-        f"    {BEGIN_SEARCH_RESULT}\nThe director of Jaws is Steven Spielberg...\n{END_SEARCH_RESULT}\n\n"
-        "Continues reasoning with the new information.\n"
-        "Assistant:\n"
-        f"    {BEGIN_SEARCH_QUERY}Where is Steven Spielberg from?{END_SEARCH_QUERY}\n\n"
-        "User:\n"
-        f"    {BEGIN_SEARCH_RESULT}\nSteven Allan Spielberg is an American filmmaker...\n{END_SEARCH_RESULT}\n\n"
-        "Continues reasoning with the new information...\n\n"
-        "Assistant:\n"
-        f"    {BEGIN_SEARCH_QUERY}Who is the director of Casino Royale?{END_SEARCH_QUERY}\n\n"
-        "User:\n"
-        f"    {BEGIN_SEARCH_RESULT}\nCasino Royale is a 2006 spy film directed by Martin Campbell...\n{END_SEARCH_RESULT}\n\n"
-        "Continues reasoning with the new information...\n\n"
-        "Assistant:\n"
-        f"    {BEGIN_SEARCH_QUERY}Where is Martin Campbell from?{END_SEARCH_QUERY}\n\n"
-        "User:\n"
-        f"    {BEGIN_SEARCH_RESULT}\nMartin Campbell (born 24 October 1943) is a New Zealand film and television director...\n{END_SEARCH_RESULT}\n\n"
-        "Continues reasoning with the new information...\n\n"
-        "Assistant:\nIt's enough to answer the question\n"
+REASON_PROMPT = f"""You are an advanced reasoning agent. Your goal is to answer the user's question by breaking it down into a series of verifiable steps.
 
-        "-- Example 2 --\n" #########################################
-        "Question: \"When was the founder of craigslist born?\"\n"
-        "Assistant:\n"
-        f"    {BEGIN_SEARCH_QUERY}Who was the founder of craigslist?{END_SEARCH_QUERY}\n\n"
-        "User:\n"
-        f"    {BEGIN_SEARCH_RESULT}\nCraigslist was founded by Craig Newmark...\n{END_SEARCH_RESULT}\n\n"
-        "Continues reasoning with the new information.\n"
-        "Assistant:\n"
-        f"    {BEGIN_SEARCH_QUERY} When was Craig Newmark born?{END_SEARCH_QUERY}\n\n"
-        "User:\n"
-        f"    {BEGIN_SEARCH_RESULT}\nCraig Newmark was born on December 6, 1952...\n{END_SEARCH_RESULT}\n\n"
-        "Continues reasoning with the new information...\n\n"
-        "Assistant:\nIt's enough to answer the question\n"
-        "**Remember**:\n"
-        f"- You have a dataset to search, so you just provide a proper search query.\n"
-        f"- Use {BEGIN_SEARCH_QUERY} to request a dataset search and end with {END_SEARCH_QUERY}.\n"
-        "- The language of query MUST be as the same as 'Question' or 'search result'.\n"
-        "- If no helpful information can be found, rewrite the search query to be less and precise keywords.\n"
-        "- When done searching, continue your reasoning.\n\n"
-        'Please answer the following question. You should think step by step to solve it.\n\n'
-    )
+You have access to a powerful search tool to find information.
 
-RELEVANT_EXTRACTION_PROMPT = """**Task Instruction:**
+**Your Task:**
+1.  Analyze the user's question.
+2.  If you need information, issue a search query to find a specific fact.
+3.  Review the search results.
+4.  Repeat the search process until you have all the facts needed to answer the question.
+5.  Once you have gathered sufficient information, synthesize the facts and provide the final answer directly.
 
-    You are tasked with reading and analyzing web pages based on the following inputs: **Previous Reasoning Steps**, **Current Search Query**, and **Searched Web Pages**. Your objective is to extract relevant and helpful information for **Current Search Query** from the **Searched Web Pages** and seamlessly integrate this information into the **Previous Reasoning Steps** to continue reasoning for the original question.
+**Tool Usage:**
+- To search, you MUST write your query between the special tokens: {BEGIN_SEARCH_QUERY}your query{END_SEARCH_QUERY}.
+- The system will provide results between {BEGIN_SEARCH_RESULT}search results{END_SEARCH_RESULT}.
+- You have a maximum of {MAX_SEARCH_LIMIT} search attempts.
 
-    **Guidelines:**
+---
+**Example 1: Multi-hop Question**
 
-    1. **Analyze the Searched Web Pages:**
-    - Carefully review the content of each searched web page.
-    - Identify factual information that is relevant to the **Current Search Query** and can aid in the reasoning process for the original question.
+**Question:** "Are both the directors of Jaws and Casino Royale from the same country?"
 
-    2. **Extract Relevant Information:**
-    - Select the information from the Searched Web Pages that directly contributes to advancing the **Previous Reasoning Steps**.
-    - Ensure that the extracted information is accurate and relevant.
+**Your Thought Process & Actions:**
+First, I need to identify the director of Jaws.
+{BEGIN_SEARCH_QUERY}who is the director of Jaws?{END_SEARCH_QUERY}
+[System returns search results]
+{BEGIN_SEARCH_RESULT}
+Jaws is a 1975 American thriller film directed by Steven Spielberg.
+{END_SEARCH_RESULT}
+Okay, the director of Jaws is Steven Spielberg. Now I need to find out his nationality.
+{BEGIN_SEARCH_QUERY}where is Steven Spielberg from?{END_SEARCH_QUERY}
+[System returns search results]
+{BEGIN_SEARCH_RESULT}
+Steven Allan Spielberg is an American filmmaker. Born in Cincinnati, Ohio...
+{END_SEARCH_RESULT}
+So, Steven Spielberg is from the USA. Next, I need to find the director of Casino Royale.
+{BEGIN_SEARCH_QUERY}who is the director of Casino Royale 2006?{END_SEARCH_QUERY}
+[System returns search results]
+{BEGIN_SEARCH_RESULT}
+Casino Royale is a 2006 spy film directed by Martin Campbell.
+{END_SEARCH_RESULT}
+The director of Casino Royale is Martin Campbell. Now I need his nationality.
+{BEGIN_SEARCH_QUERY}where is Martin Campbell from?{END_SEARCH_QUERY}
+[System returns search results]
+{BEGIN_SEARCH_RESULT}
+Martin Campbell (born 24 October 1943) is a New Zealand film and television director.
+{END_SEARCH_RESULT}
+I have all the information. Steven Spielberg is from the USA, and Martin Campbell is from New Zealand. They are not from the same country.
 
-    3. **Output Format:**
-    - **If the web pages provide helpful information for current search query:** Present the information beginning with `**Final Information**` as shown below.
-    - The language of query **MUST BE** as the same as 'Search Query' or 'Web Pages'.\n"
-    **Final Information**
+Final Answer: No, the directors of Jaws and Casino Royale are not from the same country. Steven Spielberg is from the USA, and Martin Campbell is from New Zealand.
 
-    [Helpful information]
+---
+**Example 2: Simple Fact Retrieval**
 
-    - **If the web pages do not provide any helpful information for current search query:** Output the following text.
+**Question:** "When was the founder of craigslist born?"
 
-    **Final Information**
+**Your Thought Process & Actions:**
+First, I need to know who founded craigslist.
+{BEGIN_SEARCH_QUERY}who founded craigslist?{END_SEARCH_QUERY}
+[System returns search results]
+{BEGIN_SEARCH_RESULT}
+Craigslist was founded in 1995 by Craig Newmark.
+{END_SEARCH_RESULT}
+The founder is Craig Newmark. Now I need his birth date.
+{BEGIN_SEARCH_QUERY}when was Craig Newmark born?{END_SEARCH_QUERY}
+[System returns search results]
+{BEGIN_SEARCH_RESULT}
+Craig Newmark was born on December 6, 1952.
+{END_SEARCH_RESULT}
+I have found the answer.
 
-    No helpful information found.
+Final Answer: The founder of craigslist, Craig Newmark, was born on December 6, 1952.
 
-    **Inputs:**
-    - **Previous Reasoning Steps:**  
-    {prev_reasoning}
+---
+**Important Rules:**
+- **One Fact at a Time:** Decompose the problem and issue one search query at a time to find a single, specific piece of information.
+- **Be Precise:** Formulate clear and precise search queries. If a search fails, rephrase it.
+- **Synthesize at the End:** Do not provide the final answer until you have completed all necessary searches.
+- **Language Consistency:** Your search queries should be in the same language as the user's question.
 
-    - **Current Search Query:**  
-    {search_query}
+Now, begin your work. Please answer the following question by thinking step-by-step.
+"""
 
-    - **Searched Web Pages:**  
-    {document}
+RELEVANT_EXTRACTION_PROMPT = """You are a highly efficient information extraction module. Your sole purpose is to extract the single most relevant piece of information from the provided `Searched Web Pages` that directly answers the `Current Search Query`.
 
-    """
+**Your Task:**
+1.  Read the `Current Search Query` to understand what specific information is needed.
+2.  Scan the `Searched Web Pages` to find the answer to that query.
+3.  Extract only the essential, factual information that answers the query. Be concise.
+
+**Context (For Your Information Only):**
+The `Previous Reasoning Steps` are provided to give you context on the overall goal, but your primary focus MUST be on answering the `Current Search Query`. Do not use information from the previous steps in your output.
+
+**Output Format:**
+Your response must follow one of two formats precisely.
+
+1.  **If a direct and relevant answer is found:**
+    - Start your response immediately with `Final Information`.
+    - Provide only the extracted fact(s). Do not add any extra conversational text.
+
+    *Example:*
+    `Current Search Query`: Where is Martin Campbell from?
+    `Searched Web Pages`: [Long article snippet about Martin Campbell's career, which includes the sentence "Martin Campbell (born 24 October 1943) is a New Zealand film and television director..."]
+    
+    *Your Output:*
+    Final Information
+    Martin Campbell is a New Zealand film and television director.
+
+2.  **If no relevant answer that directly addresses the query is found in the web pages:**
+    - Start your response immediately with `Final Information`.
+    - Write the exact phrase: `No helpful information found.`
+
+---
+**BEGIN TASK**
+
+**Inputs:**
+
+- **Previous Reasoning Steps:**
+{prev_reasoning}
+
+- **Current Search Query:**
+{search_query}
+
+- **Searched Web Pages:**
+{document}
+"""

api/apps/conversation_app.py

@@ -66,7 +66,8 @@ def set_conversation():
         e, dia = DialogService.get_by_id(req["dialog_id"])
         if not e:
             return get_data_error_result(message="Dialog not found")
-        conv = {"id": conv_id, "dialog_id": req["dialog_id"], "name": name, "message": [{"role": "assistant", "content": dia.prompt_config["prologue"]}],"user_id": current_user.id}
+        conv = {"id": conv_id, "dialog_id": req["dialog_id"], "name": name, "message": [{"role": "assistant", "content": dia.prompt_config["prologue"]}],"user_id": current_user.id,
+                "reference":[{}],}
         ConversationService.save(**conv)
         return get_json_result(data=conv)
     except Exception as e:

api/apps/dialog_app.py

@@ -32,7 +32,8 @@ from api.utils.api_utils import get_json_result
 @login_required
 def set_dialog():
     req = request.json
-    dialog_id = req.get("dialog_id")
+    dialog_id = req.get("dialog_id", "")
+    is_create = not dialog_id
     name = req.get("name", "New Dialog")
     if not isinstance(name, str):
         return get_data_error_result(message="Dialog name must be string.")
@@ -52,15 +53,16 @@ def set_dialog():
     llm_setting = req.get("llm_setting", {})
     prompt_config = req["prompt_config"]
 
-    if not req.get("kb_ids", []) and not prompt_config.get("tavily_api_key") and "{knowledge}" in prompt_config['system']:
-        return get_data_error_result(message="Please remove `{knowledge}` in system prompt since no knowledge base/Tavily used here.")
+    if not is_create:
+        if not req.get("kb_ids", []) and not prompt_config.get("tavily_api_key") and "{knowledge}" in prompt_config['system']:
+            return get_data_error_result(message="Please remove `{knowledge}` in system prompt since no knowledge base/Tavily used here.")
 
-    for p in prompt_config["parameters"]:
-        if p["optional"]:
-            continue
-        if prompt_config["system"].find("{%s}" % p["key"]) < 0:
-            return get_data_error_result(
-                message="Parameter '{}' is not used".format(p["key"]))
+        for p in prompt_config["parameters"]:
+            if p["optional"]:
+                continue
+            if prompt_config["system"].find("{%s}" % p["key"]) < 0:
+                return get_data_error_result(
+                    message="Parameter '{}' is not used".format(p["key"]))
 
     try:
         e, tenant = TenantService.get_by_id(current_user.id)
@@ -153,6 +155,43 @@ def list_dialogs():
         return server_error_response(e)
 
 
+@manager.route('/next', methods=['POST'])  # noqa: F821
+@login_required
+def list_dialogs_next():
+    keywords = request.args.get("keywords", "")
+    page_number = int(request.args.get("page", 0))
+    items_per_page = int(request.args.get("page_size", 0))
+    parser_id = request.args.get("parser_id")
+    orderby = request.args.get("orderby", "create_time")
+    if request.args.get("desc", "true").lower() == "false":
+        desc = False
+    else:
+        desc = True
+
+    req = request.get_json()
+    owner_ids = req.get("owner_ids", [])
+    try:
+        if not owner_ids:
+            # tenants = TenantService.get_joined_tenants_by_user_id(current_user.id)
+            # tenants = [tenant["tenant_id"] for tenant in tenants]
+            tenants = [] # keep it here
+            dialogs, total = DialogService.get_by_tenant_ids(
+                tenants, current_user.id, page_number,
+                items_per_page, orderby, desc, keywords, parser_id)
+        else:
+            tenants = owner_ids
+            dialogs, total = DialogService.get_by_tenant_ids(
+                tenants, current_user.id, 0,
+                0, orderby, desc, keywords, parser_id)
+            dialogs = [dialog for dialog in dialogs if dialog["tenant_id"] in tenants]
+            total = len(dialogs)
+            if page_number and items_per_page:
+                dialogs = dialogs[(page_number-1)*items_per_page:page_number*items_per_page]
+        return get_json_result(data={"dialogs": dialogs, "total": total})
+    except Exception as e:
+        return server_error_response(e)
+
+
 @manager.route('/rm', methods=['POST'])  # noqa: F821
 @login_required
 @validate_request("dialog_ids")

api/apps/document_app.py

@@ -206,6 +206,8 @@ def list_docs():
         desc = False
     else:
         desc = True
+    create_time_from = int(request.args.get("create_time_from", 0))
+    create_time_to = int(request.args.get("create_time_to", 0))
 
     req = request.get_json()
 
@@ -226,6 +228,14 @@ def list_docs():
     try:
         docs, tol = DocumentService.get_by_kb_id(kb_id, page_number, items_per_page, orderby, desc, keywords, run_status, types, suffix)
 
+        if create_time_from or create_time_to:
+            filtered_docs = []
+            for doc in docs:
+                doc_create_time = doc.get("create_time", 0)
+                if (create_time_from == 0 or doc_create_time >= create_time_from) and (create_time_to == 0 or doc_create_time <= create_time_to):
+                    filtered_docs.append(doc)
+            docs = filtered_docs
+
         for doc_item in docs:
             if doc_item["thumbnail"] and not doc_item["thumbnail"].startswith(IMG_BASE64_PREFIX):
                 doc_item["thumbnail"] = f"/v1/document/image/{kb_id}-{doc_item['thumbnail']}"

api/apps/kb_app.py

@@ -247,7 +247,10 @@ def list_tags(kb_id):
             code=settings.RetCode.AUTHENTICATION_ERROR
         )
 
-    tags = settings.retrievaler.all_tags(current_user.id, [kb_id])
+    tenants = UserTenantService.get_tenants_by_user_id(current_user.id)
+    tags = []
+    for tenant in tenants:
+        tags += settings.retrievaler.all_tags(tenant["tenant_id"], [kb_id])
     return get_json_result(data=tags)
 
 
@@ -263,7 +266,10 @@ def list_tags_from_kbs():
                 code=settings.RetCode.AUTHENTICATION_ERROR
             )
 
-    tags = settings.retrievaler.all_tags(current_user.id, kb_ids)
+    tenants = UserTenantService.get_tenants_by_user_id(current_user.id)
+    tags = []
+    for tenant in tenants:
+        tags += settings.retrievaler.all_tags(tenant["tenant_id"], kb_ids)
     return get_json_result(data=tags)
 
 

api/apps/llm_app.py

@@ -15,7 +15,6 @@
 #
 import logging
 import json
-import base64
 from flask import request
 from flask_login import login_required, current_user
 from api.db.services.llm_service import LLMFactoriesService, TenantLLMService, LLMService
@@ -24,7 +23,7 @@ from api.utils.api_utils import server_error_response, get_data_error_result, va
 from api.db import StatusEnum, LLMType
 from api.db.db_models import TenantLLM
 from api.utils.api_utils import get_json_result
-from api.utils.base64_image import test_image_base64
+from api.utils.base64_image import test_image
 from rag.llm import EmbeddingModel, ChatModel, RerankModel, CvModel, TTSModel
 
 
@@ -256,7 +255,7 @@ def add_llm():
             base_url=llm["api_base"]
         )
         try:
-            image_data = base64.b64decode(test_image_base64)
+            image_data = test_image
             m, tc = mdl.describe(image_data)
             if not m and not tc:
                 raise Exception(m)

api/apps/sdk/dify_retrieval.py

@@ -1,4 +1,4 @@
-#
+    #
 #  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
 #
 #  Licensed under the Apache License, Version 2.0 (the "License");
@@ -13,6 +13,8 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import logging
+
 from flask import request, jsonify
 
 from api.db import LLMType
@@ -73,11 +75,13 @@ def retrieval(tenant_id):
         for c in ranks["chunks"]:
             e, doc = DocumentService.get_by_id( c["doc_id"])
             c.pop("vector", None)
+            meta = getattr(doc, 'meta_fields', {})
+            meta["doc_id"] = c["doc_id"]
             records.append({
                 "content": c["content_with_weight"],
                 "score": c["similarity"],
                 "title": c["docnm_kwd"],
-                "metadata": doc.meta_fields
+                "metadata": meta
             })
 
         return jsonify({"records": records})
@@ -87,4 +91,5 @@ def retrieval(tenant_id):
                 message='No chunk found! Check the chunk status please!',
                 code=settings.RetCode.NOT_FOUND
             )
+        logging.exception(e)
         return build_error_result(message=str(e), code=settings.RetCode.SERVER_ERROR)

api/apps/sdk/doc.py

@@ -38,7 +38,7 @@ from api.utils.api_utils import check_duplicate_ids, construct_json_result, get_
 from rag.app.qa import beAdoc, rmPrefix
 from rag.app.tag import label_question
 from rag.nlp import rag_tokenizer, search
-from rag.prompts import keyword_extraction, cross_languages
+from rag.prompts import cross_languages, keyword_extraction
 from rag.utils import rmSpace
 from rag.utils.storage_factory import STORAGE_IMPL
 
@@ -456,6 +456,18 @@ def list_docs(dataset_id, tenant_id):
         required: false
         default: true
         description: Order in descending.
+    - in: query
+        name: create_time_from
+        type: integer
+        required: false
+        default: 0
+        description: Unix timestamp for filtering documents created after this time. 0 means no filter.
+      - in: query
+        name: create_time_to
+        type: integer
+        required: false
+        default: 0
+        description: Unix timestamp for filtering documents created before this time. 0 means no filter.
       - in: header
         name: Authorization
         type: string
@@ -517,6 +529,17 @@ def list_docs(dataset_id, tenant_id):
         desc = True
     docs, tol = DocumentService.get_list(dataset_id, page, page_size, orderby, desc, keywords, id, name)
 
+    create_time_from = int(request.args.get("create_time_from", 0))
+    create_time_to = int(request.args.get("create_time_to", 0))
+
+    if create_time_from or create_time_to:
+        filtered_docs = []
+        for doc in docs:
+            doc_create_time = doc.get("create_time", 0)
+            if (create_time_from == 0 or doc_create_time >= create_time_from) and (create_time_to == 0 or doc_create_time <= create_time_to):
+                filtered_docs.append(doc)
+        docs = filtered_docs
+
     # rename key's name
     renamed_doc_list = []
     key_mapping = {

api/apps/sdk/session.py

@@ -51,6 +51,7 @@ def create(tenant_id, chat_id):
         "name": req.get("name", "New session"),
         "message": [{"role": "assistant", "content": dia[0].prompt_config.get("prologue")}],
         "user_id": req.get("user_id", ""),
+        "reference":[{}],
     }
     if not conv.get("name"):
         return get_error_data_result(message="`name` can not be empty.")
@@ -435,14 +436,38 @@ def agents_completion_openai_compatibility(tenant_id, agent_id):
             )
         )
 
-    # Get the last user message as the question
     question = next((m["content"] for m in reversed(messages) if m["role"] == "user"), "")
 
-    if req.get("stream", True):
-        return Response(completionOpenAI(tenant_id, agent_id, question, session_id=req.get("id", req.get("metadata", {}).get("id", "")), stream=True), mimetype="text/event-stream")
+    stream = req.pop("stream", False)
+    if stream:
+        resp = Response(
+            completionOpenAI(
+                tenant_id,
+                agent_id,
+                question,
+                session_id=req.get("id", req.get("metadata", {}).get("id", "")),
+                stream=True,
+                **req,
+            ),
+            mimetype="text/event-stream",
+        )
+        resp.headers.add_header("Cache-control", "no-cache")
+        resp.headers.add_header("Connection", "keep-alive")
+        resp.headers.add_header("X-Accel-Buffering", "no")
+        resp.headers.add_header("Content-Type", "text/event-stream; charset=utf-8")
+        return resp
     else:
         # For non-streaming, just return the response directly
-        response = next(completionOpenAI(tenant_id, agent_id, question, session_id=req.get("id", req.get("metadata", {}).get("id", "")), stream=False))
+        response = next(
+            completionOpenAI(
+                tenant_id,
+                agent_id,
+                question,
+                session_id=req.get("id", req.get("metadata", {}).get("id", "")),
+                stream=False,
+                **req,
+            )
+        )
         return jsonify(response)
 
 
@@ -512,16 +537,16 @@ def list_session(tenant_id, chat_id):
             if "prompt" in info:
                 info.pop("prompt")
         conv["chat_id"] = conv.pop("dialog_id")
-        if conv["reference"]:
+        ref_messages = conv["reference"]
+        if ref_messages:
             messages = conv["messages"]
             message_num = 0
-            while message_num < len(messages) and message_num < len(conv["reference"]):
-                if message_num != 0 and messages[message_num]["role"] != "user":
-                    if message_num >= len(conv["reference"]):
-                        break
+            ref_num = 0
+            while message_num < len(messages) and ref_num < len(ref_messages):
+                if messages[message_num]["role"] != "user":
                     chunk_list = []
-                    if "chunks" in conv["reference"][message_num]:
-                        chunks = conv["reference"][message_num]["chunks"]
+                    if "chunks" in ref_messages[ref_num]:
+                        chunks = ref_messages[ref_num]["chunks"]
                         for chunk in chunks:
                             new_chunk = {
                                 "id": chunk.get("chunk_id", chunk.get("id")),
@@ -535,6 +560,7 @@ def list_session(tenant_id, chat_id):
 
                             chunk_list.append(new_chunk)
                     messages[message_num]["reference"] = chunk_list
+                    ref_num += 1
                 message_num += 1
         del conv["reference"]
     return get_result(data=convs)
@@ -848,10 +874,10 @@ def begin_inputs(agent_id):
         return get_error_data_result(f"Can't find agent by ID: {agent_id}")
 
     canvas = Canvas(json.dumps(cvs.dsl), objs[0].tenant_id)
-    return get_result(data={
-        "title": cvs.title,
-        "avatar": cvs.avatar,
-        "inputs": canvas.get_component_input_form("begin")
-    })
-
-
+    return get_result(
+        data={
+            "title": cvs.title,
+            "avatar": cvs.avatar,
+            "inputs": canvas.get_component_input_form("begin"),
+        }
+    )

api/db/services/canvas_service.py

@@ -16,7 +16,6 @@
 import json
 import logging
 import time
-import traceback
 from uuid import uuid4
 from agent.canvas import Canvas
 from api.db import TenantPermission
@@ -54,12 +53,12 @@ class UserCanvasService(CommonService):
         agents = agents.paginate(page_number, items_per_page)
 
         return list(agents.dicts())
-   
+
     @classmethod
     @DB.connection_context()
     def get_by_tenant_id(cls, pid):
         try:
-            
+
             fields = [
                 cls.model.id,
                 cls.model.avatar,
@@ -83,7 +82,7 @@ class UserCanvasService(CommonService):
         except Exception as e:
             logging.exception(e)
             return False, None
-     
+
     @classmethod
     @DB.connection_context()
     def get_by_tenant_ids(cls, joined_tenant_ids, user_id,
@@ -103,14 +102,14 @@ class UserCanvasService(CommonService):
         ]
         if keywords:
             agents = cls.model.select(*fields).join(User, on=(cls.model.user_id == User.id)).where(
-                ((cls.model.user_id.in_(joined_tenant_ids) & (cls.model.permission == 
+                ((cls.model.user_id.in_(joined_tenant_ids) & (cls.model.permission ==
                                                                 TenantPermission.TEAM.value)) | (
                     cls.model.user_id == user_id)),
                 (fn.LOWER(cls.model.title).contains(keywords.lower()))
             )
         else:
             agents = cls.model.select(*fields).join(User, on=(cls.model.user_id == User.id)).where(
-                ((cls.model.user_id.in_(joined_tenant_ids) & (cls.model.permission == 
+                ((cls.model.user_id.in_(joined_tenant_ids) & (cls.model.permission ==
                                                                 TenantPermission.TEAM.value)) | (
                     cls.model.user_id == user_id))
             )
@@ -173,223 +172,104 @@ def completion(tenant_id, agent_id, session_id=None, **kwargs):
     conv.message.append({"role": "assistant", "content": txt, "created_at": time.time(), "id": message_id})
     conv.reference = canvas.get_reference()
     conv.errors = canvas.error
-    API4ConversationService.append_message(conv.id, conv.to_dict())
+    conv = conv.to_dict()
+    API4ConversationService.append_message(conv["id"], conv)
 
 
 def completionOpenAI(tenant_id, agent_id, question, session_id=None, stream=True, **kwargs):
-    """Main function for OpenAI-compatible completions, structured similarly to the completion function."""
     tiktokenenc = tiktoken.get_encoding("cl100k_base")
-    e, cvs = UserCanvasService.get_by_id(agent_id)
-    
-    if not e:
-        yield get_data_openai(
-            id=session_id,
-            model=agent_id,
-            content="**ERROR**: Agent not found."
-        )
-        return
-    
-    if cvs.user_id != tenant_id:
-        yield get_data_openai(
-            id=session_id,
-            model=agent_id,
-            content="**ERROR**: You do not own the agent"
-        )
-        return
-    
-    if not isinstance(cvs.dsl, str):
-        cvs.dsl = json.dumps(cvs.dsl, ensure_ascii=False)
-    
-    canvas = Canvas(cvs.dsl, tenant_id)
-    canvas.reset()
-    message_id = str(uuid4())
-    
-    # Handle new session creation
-    if not session_id:
-        query = canvas.get_preset_param()
-        if query:
-            for ele in query:
-                if not ele["optional"]:
-                    if not kwargs.get(ele["key"]):
-                        yield get_data_openai(
-                            id=None,
-                            model=agent_id,
-                            content=f"`{ele['key']}` is required",
-                            completion_tokens=len(tiktokenenc.encode(f"`{ele['key']}` is required")),
-                            prompt_tokens=len(tiktokenenc.encode(question if question else ""))
-                        )
-                        return
-                    ele["value"] = kwargs[ele["key"]]
-                if ele["optional"]:
-                    if kwargs.get(ele["key"]):
-                        ele["value"] = kwargs[ele['key']]
-                    else:
-                        if "value" in ele:
-                            ele.pop("value")
-        
-        cvs.dsl = json.loads(str(canvas))
-        session_id = get_uuid()
-        conv = {
-            "id": session_id,
-            "dialog_id": cvs.id,
-            "user_id": kwargs.get("user_id", "") if isinstance(kwargs, dict) else "",
-            "message": [{"role": "assistant", "content": canvas.get_prologue(), "created_at": time.time()}],
-            "source": "agent",
-            "dsl": cvs.dsl
-        }
-        canvas.messages.append({"role": "user", "content": question, "id": message_id})
-        canvas.add_user_input(question)
-
-        API4ConversationService.save(**conv)
-        conv = API4Conversation(**conv)
-        if not conv.message:
-            conv.message = []
-        conv.message.append({
-            "role": "user",
-            "content": question,
-            "id": message_id
-        })
-        
-        if not conv.reference:
-            conv.reference = []
-        conv.reference.append({"chunks": [], "doc_aggs": []})
-            
-    # Handle existing session
-    else:
-        e, conv = API4ConversationService.get_by_id(session_id)
-        if not e:
-            yield get_data_openai(
-                id=session_id,
-                model=agent_id,
-                content="**ERROR**: Session not found!"
-            )
-            return
-        
-        canvas = Canvas(json.dumps(conv.dsl), tenant_id)
-        canvas.messages.append({"role": "user", "content": question, "id": message_id})
-        canvas.add_user_input(question)
-        
-        if not conv.message:
-            conv.message = []
-        conv.message.append({
-            "role": "user",
-            "content": question,
-            "id": message_id
-        })
-        
-        if not conv.reference:
-            conv.reference = []
-        conv.reference.append({"chunks": [], "doc_aggs": []})
-    
-    # Process request based on stream mode
-    final_ans = {"reference": [], "content": ""}
     prompt_tokens = len(tiktokenenc.encode(str(question)))
-    
+    user_id = kwargs.get("user_id", "")
+
     if stream:
+        completion_tokens = 0
         try:
-            completion_tokens = 0
-            for ans in canvas.run(stream=True, bypass_begin=True):
-                if ans.get("running_status"):
-                    completion_tokens += len(tiktokenenc.encode(ans.get("content", "")))
-                    yield "data: " + json.dumps(
-                        get_data_openai(
-                            id=session_id,
-                            model=agent_id,
-                            content=ans["content"],
-                            object="chat.completion.chunk",
-                            completion_tokens=completion_tokens,
-                            prompt_tokens=prompt_tokens
-                        ),
-                        ensure_ascii=False
-                    ) + "\n\n"
+            for ans in completion(
+                tenant_id=tenant_id,
+                agent_id=agent_id,
+                session_id=session_id,
+                query=question,
+                user_id=user_id,
+                **kwargs
+            ):
+                if isinstance(ans, str):
+                    try:
+                        ans = json.loads(ans[5:])  # remove "data:"
+                    except Exception as e:
+                        logging.exception(f"Agent OpenAI-Compatible completionOpenAI parse answer failed: {e}")
+                        continue
+
+                if ans.get("event") != "message":
                     continue
-                
-                for k in ans.keys():
-                    final_ans[k] = ans[k]
-                
-                completion_tokens += len(tiktokenenc.encode(final_ans.get("content", "")))
+
+                content_piece = ans["data"]["content"]
+                completion_tokens += len(tiktokenenc.encode(content_piece))
+
                 yield "data: " + json.dumps(
                     get_data_openai(
-                        id=session_id,
+                        id=session_id or str(uuid4()),
                         model=agent_id,
-                        content=final_ans["content"],
-                        object="chat.completion.chunk",
-                        finish_reason="stop",
+                        content=content_piece,
+                        prompt_tokens=prompt_tokens,
                         completion_tokens=completion_tokens,
-                        prompt_tokens=prompt_tokens
+                        stream=True
                     ),
                     ensure_ascii=False
                 ) + "\n\n"
-            
-            # Update conversation
-            canvas.messages.append({"role": "assistant", "content": final_ans["content"], "created_at": time.time(), "id": message_id})
-            canvas.history.append(("assistant", final_ans["content"]))
-            if final_ans.get("reference"):
-                canvas.reference.append(final_ans["reference"])
-            conv.dsl = json.loads(str(canvas))
-            API4ConversationService.append_message(conv.id, conv.to_dict())
-            
+
             yield "data: [DONE]\n\n"
-            
+
         except Exception as e:
-            traceback.print_exc()
-            conv.dsl = json.loads(str(canvas))
-            API4ConversationService.append_message(conv.id, conv.to_dict())
             yield "data: " + json.dumps(
                 get_data_openai(
-                    id=session_id,
+                    id=session_id or str(uuid4()),
                     model=agent_id,
-                    content="**ERROR**: " + str(e),
+                    content=f"**ERROR**: {str(e)}",
                     finish_reason="stop",
-                    completion_tokens=len(tiktokenenc.encode("**ERROR**: " + str(e))),
-                    prompt_tokens=prompt_tokens
+                    prompt_tokens=prompt_tokens,
+                    completion_tokens=len(tiktokenenc.encode(f"**ERROR**: {str(e)}")),
+                    stream=True
                 ),
                 ensure_ascii=False
             ) + "\n\n"
             yield "data: [DONE]\n\n"
-    
-    else:  # Non-streaming mode
+
+    else:
         try:
-            all_answer_content = ""
-            for answer in canvas.run(stream=False, bypass_begin=True):
-                if answer.get("running_status"):
+            all_content = ""
+            for ans in completion(
+                tenant_id=tenant_id,
+                agent_id=agent_id,
+                session_id=session_id,
+                query=question,
+                user_id=user_id,
+                **kwargs
+            ):
+                if isinstance(ans, str):
+                    ans = json.loads(ans[5:])
+                if ans.get("event") != "message":
                     continue
-                
-                final_ans["content"] = "\n".join(answer["content"]) if "content" in answer else ""
-                final_ans["reference"] = answer.get("reference", [])
-                all_answer_content += final_ans["content"]
-            
-            final_ans["content"] = all_answer_content
-            
-            # Update conversation
-            canvas.messages.append({"role": "assistant", "content": final_ans["content"], "created_at": time.time(), "id": message_id})
-            canvas.history.append(("assistant", final_ans["content"]))
-            if final_ans.get("reference"):
-                canvas.reference.append(final_ans["reference"])
-            conv.dsl = json.loads(str(canvas))
-            API4ConversationService.append_message(conv.id, conv.to_dict())
-            
-            # Return the response in OpenAI format
+                all_content += ans["data"]["content"]
+
+            completion_tokens = len(tiktokenenc.encode(all_content))
+
             yield get_data_openai(
-                id=session_id,
+                id=session_id or str(uuid4()),
                 model=agent_id,
-                content=final_ans["content"],
-                finish_reason="stop",
-                completion_tokens=len(tiktokenenc.encode(final_ans["content"])),
                 prompt_tokens=prompt_tokens,
-                param=canvas.get_preset_param()  # Added param info like in completion
-            )
-            
-        except Exception as e:
-            traceback.print_exc()
-            conv.dsl = json.loads(str(canvas))
-            API4ConversationService.append_message(conv.id, conv.to_dict())
-            yield get_data_openai(
-                id=session_id,
-                model=agent_id,
-                content="**ERROR**: " + str(e),
+                completion_tokens=completion_tokens,
+                content=all_content,
                 finish_reason="stop",
-                completion_tokens=len(tiktokenenc.encode("**ERROR**: " + str(e))),
-                prompt_tokens=prompt_tokens
+                param=None
             )
 
+        except Exception as e:
+            yield get_data_openai(
+                id=session_id or str(uuid4()),
+                model=agent_id,
+                prompt_tokens=prompt_tokens,
+                completion_tokens=len(tiktokenenc.encode(f"**ERROR**: {str(e)}")),
+                content=f"**ERROR**: {str(e)}",
+                finish_reason="stop",
+                param=None
+            )

api/db/services/dialog_service.py

@@ -23,6 +23,7 @@ from functools import partial
 from timeit import default_timer as timer
 
 from langfuse import Langfuse
+from peewee import fn
 
 from agentic_reasoning import DeepResearcher
 from api import settings
@@ -96,6 +97,66 @@ class DialogService(CommonService):
         return list(chats.dicts())
 
 
+    @classmethod
+    @DB.connection_context()
+    def get_by_tenant_ids(cls, joined_tenant_ids, user_id, page_number, items_per_page, orderby, desc, keywords, parser_id=None):
+        from api.db.db_models import User
+
+        fields = [
+            cls.model.id,
+            cls.model.tenant_id,
+            cls.model.name,
+            cls.model.description,
+            cls.model.language,
+            cls.model.llm_id,
+            cls.model.llm_setting,
+            cls.model.prompt_type,
+            cls.model.prompt_config,
+            cls.model.similarity_threshold,
+            cls.model.vector_similarity_weight,
+            cls.model.top_n,
+            cls.model.top_k,
+            cls.model.do_refer,
+            cls.model.rerank_id,
+            cls.model.kb_ids,
+            cls.model.status,
+            User.nickname,
+            User.avatar.alias("tenant_avatar"),
+            cls.model.update_time,
+            cls.model.create_time,
+        ]
+        if keywords:
+            dialogs = (
+                cls.model.select(*fields)
+                .join(User, on=(cls.model.tenant_id == User.id))
+                .where(
+                    (cls.model.tenant_id.in_(joined_tenant_ids) | (cls.model.tenant_id == user_id)) & (cls.model.status == StatusEnum.VALID.value),
+                    (fn.LOWER(cls.model.name).contains(keywords.lower())),
+                )
+            )
+        else:
+            dialogs = (
+                cls.model.select(*fields)
+                .join(User, on=(cls.model.tenant_id == User.id))
+                .where(
+                    (cls.model.tenant_id.in_(joined_tenant_ids) | (cls.model.tenant_id == user_id)) & (cls.model.status == StatusEnum.VALID.value),
+                )
+            )
+        if parser_id:
+            dialogs = dialogs.where(cls.model.parser_id == parser_id)
+        if desc:
+            dialogs = dialogs.order_by(cls.model.getter_by(orderby).desc())
+        else:
+            dialogs = dialogs.order_by(cls.model.getter_by(orderby).asc())
+
+        count = dialogs.count()
+
+        if page_number and items_per_page:
+            dialogs = dialogs.paginate(page_number, items_per_page)
+
+        return list(dialogs.dicts()), count
+
+
 def chat_solo(dialog, messages, stream=True):
     if TenantLLMService.llm_id2llm_type(dialog.llm_id) == "image2text":
         chat_mdl = LLMBundle(dialog.tenant_id, LLMType.IMAGE2TEXT, dialog.llm_id)
@@ -208,12 +269,14 @@ def chat(dialog, messages, stream=True, **kwargs):
     check_llm_ts = timer()
 
     langfuse_tracer = None
+    trace_context = {}
     langfuse_keys = TenantLangfuseService.filter_by_tenant(tenant_id=dialog.tenant_id)
     if langfuse_keys:
         langfuse = Langfuse(public_key=langfuse_keys.public_key, secret_key=langfuse_keys.secret_key, host=langfuse_keys.host)
         if langfuse.auth_check():
             langfuse_tracer = langfuse
-            langfuse.trace = langfuse_tracer.trace(name=f"{dialog.name}-{llm_model_config['llm_name']}")
+            trace_id = langfuse_tracer.create_trace_id()
+            trace_context = {"trace_id": trace_id}
 
     check_langfuse_tracer_ts = timer()
     kbs, embd_mdl, rerank_mdl, chat_mdl, tts_mdl = get_models(dialog)
@@ -400,17 +463,19 @@ def chat(dialog, messages, stream=True, **kwargs):
             f"  - Token speed: {int(tk_num / (generate_result_time_cost / 1000.0))}/s"
         )
 
-        langfuse_output = "\n" + re.sub(r"^.*?(### Query:.*)", r"\1", prompt, flags=re.DOTALL)
-        langfuse_output = {"time_elapsed:": re.sub(r"\n", "  \n", langfuse_output), "created_at": time.time()}
-
         # Add a condition check to call the end method only if langfuse_tracer exists
         if langfuse_tracer and "langfuse_generation" in locals():
-            langfuse_generation.end(output=langfuse_output)
+            langfuse_output = "\n" + re.sub(r"^.*?(### Query:.*)", r"\1", prompt, flags=re.DOTALL)
+            langfuse_output = {"time_elapsed:": re.sub(r"\n", "  \n", langfuse_output), "created_at": time.time()}
+            langfuse_generation.update(output=langfuse_output)
+            langfuse_generation.end()
 
         return {"answer": think + answer, "reference": refs, "prompt": re.sub(r"\n", "  \n", prompt), "created_at": time.time()}
 
     if langfuse_tracer:
-        langfuse_generation = langfuse_tracer.trace.generation(name="chat", model=llm_model_config["llm_name"], input={"prompt": prompt, "prompt4citation": prompt4citation, "messages": msg})
+        langfuse_generation = langfuse_tracer.start_generation(
+            trace_context=trace_context, name="chat", model=llm_model_config["llm_name"], input={"prompt": prompt, "prompt4citation": prompt4citation, "messages": msg}
+        )
 
     if stream:
         last_ans = ""

api/db/services/llm_service.py

@@ -217,7 +217,7 @@ class TenantLLMService(CommonService):
         return list(objs)
 
     @staticmethod
-    def llm_id2llm_type(llm_id: str) ->str|None:
+    def llm_id2llm_type(llm_id: str) -> str | None:
         llm_id, *_ = TenantLLMService.split_model_name_and_factory(llm_id)
         llm_factories = settings.FACTORY_LLM_INFOS
         for llm_factory in llm_factories:
@@ -225,6 +225,9 @@ class TenantLLMService(CommonService):
                 if llm_id == llm["llm_name"]:
                     return llm["model_type"].split(",")[-1]
 
+        for llm in LLMService.query(llm_name=llm_id):
+            return llm.model_type
+
 
 class LLMBundle:
     def __init__(self, tenant_id, llm_type, llm_name=None, lang="Chinese", **kwargs):
@@ -240,13 +243,13 @@ class LLMBundle:
         self.verbose_tool_use = kwargs.get("verbose_tool_use")
 
         langfuse_keys = TenantLangfuseService.filter_by_tenant(tenant_id=tenant_id)
+        self.langfuse = None
         if langfuse_keys:
             langfuse = Langfuse(public_key=langfuse_keys.public_key, secret_key=langfuse_keys.secret_key, host=langfuse_keys.host)
             if langfuse.auth_check():
                 self.langfuse = langfuse
-                self.trace = self.langfuse.trace(name=f"{self.llm_type}-{self.llm_name}")
-        else:
-            self.langfuse = None
+                trace_id = self.langfuse.create_trace_id()
+                self.trace_context = {"trace_id": trace_id}
 
     def bind_tools(self, toolcall_session, tools):
         if not self.is_tools:
@@ -256,7 +259,7 @@ class LLMBundle:
 
     def encode(self, texts: list):
         if self.langfuse:
-            generation = self.trace.generation(name="encode", model=self.llm_name, input={"texts": texts})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="encode", model=self.llm_name, input={"texts": texts})
 
         embeddings, used_tokens = self.mdl.encode(texts)
         llm_name = getattr(self, "llm_name", None)
@@ -264,13 +267,14 @@ class LLMBundle:
             logging.error("LLMBundle.encode can't update token usage for {}/EMBEDDING used_tokens: {}".format(self.tenant_id, used_tokens))
 
         if self.langfuse:
-            generation.end(usage_details={"total_tokens": used_tokens})
+            generation.update(usage_details={"total_tokens": used_tokens})
+            generation.end()
 
         return embeddings, used_tokens
 
     def encode_queries(self, query: str):
         if self.langfuse:
-            generation = self.trace.generation(name="encode_queries", model=self.llm_name, input={"query": query})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="encode_queries", model=self.llm_name, input={"query": query})
 
         emd, used_tokens = self.mdl.encode_queries(query)
         llm_name = getattr(self, "llm_name", None)
@@ -278,65 +282,70 @@ class LLMBundle:
             logging.error("LLMBundle.encode_queries can't update token usage for {}/EMBEDDING used_tokens: {}".format(self.tenant_id, used_tokens))
 
         if self.langfuse:
-            generation.end(usage_details={"total_tokens": used_tokens})
+            generation.update(usage_details={"total_tokens": used_tokens})
+            generation.end()
 
         return emd, used_tokens
 
     def similarity(self, query: str, texts: list):
         if self.langfuse:
-            generation = self.trace.generation(name="similarity", model=self.llm_name, input={"query": query, "texts": texts})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="similarity", model=self.llm_name, input={"query": query, "texts": texts})
 
         sim, used_tokens = self.mdl.similarity(query, texts)
         if not TenantLLMService.increase_usage(self.tenant_id, self.llm_type, used_tokens):
             logging.error("LLMBundle.similarity can't update token usage for {}/RERANK used_tokens: {}".format(self.tenant_id, used_tokens))
 
         if self.langfuse:
-            generation.end(usage_details={"total_tokens": used_tokens})
+            generation.update(usage_details={"total_tokens": used_tokens})
+            generation.end()
 
         return sim, used_tokens
 
     def describe(self, image, max_tokens=300):
         if self.langfuse:
-            generation = self.trace.generation(name="describe", metadata={"model": self.llm_name})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="describe", metadata={"model": self.llm_name})
 
         txt, used_tokens = self.mdl.describe(image)
         if not TenantLLMService.increase_usage(self.tenant_id, self.llm_type, used_tokens):
             logging.error("LLMBundle.describe can't update token usage for {}/IMAGE2TEXT used_tokens: {}".format(self.tenant_id, used_tokens))
 
         if self.langfuse:
-            generation.end(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.update(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.end()
 
         return txt
 
     def describe_with_prompt(self, image, prompt):
         if self.langfuse:
-            generation = self.trace.generation(name="describe_with_prompt", metadata={"model": self.llm_name, "prompt": prompt})
+            generation = self.language.start_generation(trace_context=self.trace_context, name="describe_with_prompt", metadata={"model": self.llm_name, "prompt": prompt})
 
         txt, used_tokens = self.mdl.describe_with_prompt(image, prompt)
         if not TenantLLMService.increase_usage(self.tenant_id, self.llm_type, used_tokens):
             logging.error("LLMBundle.describe can't update token usage for {}/IMAGE2TEXT used_tokens: {}".format(self.tenant_id, used_tokens))
 
         if self.langfuse:
-            generation.end(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.update(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.end()
 
         return txt
 
     def transcription(self, audio):
         if self.langfuse:
-            generation = self.trace.generation(name="transcription", metadata={"model": self.llm_name})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="transcription", metadata={"model": self.llm_name})
 
         txt, used_tokens = self.mdl.transcription(audio)
         if not TenantLLMService.increase_usage(self.tenant_id, self.llm_type, used_tokens):
             logging.error("LLMBundle.transcription can't update token usage for {}/SEQUENCE2TXT used_tokens: {}".format(self.tenant_id, used_tokens))
 
         if self.langfuse:
-            generation.end(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.update(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.end()
 
         return txt
 
     def tts(self, text: str) -> Generator[bytes, None, None]:
         if self.langfuse:
-            span = self.trace.span(name="tts", input={"text": text})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="tts", input={"text": text})
 
         for chunk in self.mdl.tts(text):
             if isinstance(chunk, int):
@@ -346,7 +355,7 @@ class LLMBundle:
             yield chunk
 
         if self.langfuse:
-            span.end()
+            generation.end()
 
     def _remove_reasoning_content(self, txt: str) -> str:
         first_think_start = txt.find("<think>")
@@ -362,9 +371,9 @@ class LLMBundle:
 
         return txt[last_think_end + len("</think>") :]
 
-    def chat(self, system: str, history: list, gen_conf: dict={}, **kwargs) -> str:
+    def chat(self, system: str, history: list, gen_conf: dict = {}, **kwargs) -> str:
         if self.langfuse:
-            generation = self.trace.generation(name="chat", model=self.llm_name, input={"system": system, "history": history})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="chat", model=self.llm_name, input={"system": system, "history": history})
 
         chat_partial = partial(self.mdl.chat, system, history, gen_conf)
         if self.is_tools and self.mdl.is_tools:
@@ -380,13 +389,14 @@ class LLMBundle:
             logging.error("LLMBundle.chat can't update token usage for {}/CHAT llm_name: {}, used_tokens: {}".format(self.tenant_id, self.llm_name, used_tokens))
 
         if self.langfuse:
-            generation.end(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.update(output={"output": txt}, usage_details={"total_tokens": used_tokens})
+            generation.end()
 
         return txt
 
-    def chat_streamly(self, system: str, history: list, gen_conf: dict={}, **kwargs):
+    def chat_streamly(self, system: str, history: list, gen_conf: dict = {}, **kwargs):
         if self.langfuse:
-            generation = self.trace.generation(name="chat_streamly", model=self.llm_name, input={"system": system, "history": history})
+            generation = self.langfuse.start_generation(trace_context=self.trace_context, name="chat_streamly", model=self.llm_name, input={"system": system, "history": history})
 
         ans = ""
         chat_partial = partial(self.mdl.chat_streamly, system, history, gen_conf)
@@ -398,7 +408,8 @@ class LLMBundle:
             if isinstance(txt, int):
                 total_tokens = txt
                 if self.langfuse:
-                    generation.end(output={"output": ans})
+                    generation.update(output={"output": ans})
+                    generation.end()
                 break
 
             if txt.endswith("</think>"):

api/settings.py

@@ -70,6 +70,7 @@ REGISTER_ENABLED = 1
 # sandbox-executor-manager
 SANDBOX_ENABLED = 0
 SANDBOX_HOST = None
+STRONG_TEST_COUNT = int(os.environ.get("STRONG_TEST_COUNT", "32"))
 
 BUILTIN_EMBEDDING_MODELS = ["BAAI/bge-large-zh-v1.5@BAAI", "maidalun1020/bce-embedding-base_v1@Youdao"]
 

api/utils/api_utils.py

@@ -402,8 +402,22 @@ def get_data_openai(
     finish_reason=None,
     object="chat.completion",
     param=None,
+    stream=False
 ):
     total_tokens = prompt_tokens + completion_tokens
+
+    if stream:
+        return {
+            "id": f"{id}",
+            "object": "chat.completion.chunk",
+            "model": model,
+            "choices": [{
+                "delta": {"content": content},
+                "finish_reason": finish_reason,
+                "index": 0,
+            }],
+        }
+
     return {
         "id": f"{id}",
         "object": object,
@@ -414,9 +428,21 @@ def get_data_openai(
             "prompt_tokens": prompt_tokens,
             "completion_tokens": completion_tokens,
             "total_tokens": total_tokens,
-            "completion_tokens_details": {"reasoning_tokens": 0, "accepted_prediction_tokens": 0, "rejected_prediction_tokens": 0},
+            "completion_tokens_details": {
+                "reasoning_tokens": 0,
+                "accepted_prediction_tokens": 0,
+                "rejected_prediction_tokens": 0,
+            },
         },
-        "choices": [{"message": {"role": "assistant", "content": content}, "logprobs": None, "finish_reason": finish_reason, "index": 0}],
+        "choices": [{
+            "message": {
+                "role": "assistant",
+                "content": content
+            },
+            "logprobs": None,
+            "finish_reason": finish_reason,
+            "index": 0,
+        }],
     }
 
 
@@ -687,7 +713,13 @@ def timeout(seconds: float | int = None, attempts: int = 2, *, exception: Option
 
 
 async def is_strong_enough(chat_model, embedding_model):
-    @timeout(30, 2)
+    count = settings.STRONG_TEST_COUNT
+    if not chat_model or not embedding_model:
+        return
+    if isinstance(count, int) and count <= 0:
+        return
+
+    @timeout(60, 2)
     async def _is_strong_enough():
         nonlocal chat_model, embedding_model
         if embedding_model:
@@ -701,5 +733,5 @@ async def is_strong_enough(chat_model, embedding_model):
 
     # Pressure test for GraphRAG task
     async with trio.open_nursery() as nursery:
-        for _ in range(32):
+        for _ in range(count):
             nursery.start_soon(_is_strong_enough)

api/utils/base64_image.py

@@ -1 +1,3 @@
+import base64
 test_image_base64 = "iVBORw0KGgoAAAANSUhEUgAAAGQAAABkCAIAAAD/gAIDAAAA6ElEQVR4nO3QwQ3AIBDAsIP9d25XIC+EZE8QZc18w5l9O+AlZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBWYFZgVmBT+IYAHHLHkdEgAAAABJRU5ErkJggg=="
+test_image = base64.b64decode(test_image_base64)

conf/llm_factories.json

@@ -2598,234 +2598,255 @@
             "tags": "LLM,TEXT EMBEDDING,TEXT RE-RANK,IMAGE2TEXT",
             "status": "1",
             "llm": [
+                {
+                    "llm_name": "Qwen3-Embedding-8B",
+                    "tags": "TEXT EMBEDDING,TEXT RE-RANK,32k",
+                    "max_tokens": 32000,
+                    "model_type": "embedding",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen3-Embedding-4B",
+                    "tags": "TEXT EMBEDDING,TEXT RE-RANK,32k",
+                    "max_tokens": 32000,
+                    "model_type": "embedding",
+                    "is_tools": false
+                },
+                {
+                    "llm_name": "Qwen3-Embedding-0.6B",
+                    "tags": "TEXT EMBEDDING,TEXT RE-RANK,32k",
+                    "max_tokens": 32000,
+                    "model_type": "embedding",
+                    "is_tools": false
+                },
                 {
                     "llm_name": "Qwen/Qwen3-235B-A22B",
                     "tags": "LLM,CHAT,128k",
-                    "max_tokens": 8192,
+                    "max_tokens": 128000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen3-30B-A3B",
                     "tags": "LLM,CHAT,128k",
-                    "max_tokens": 8192,
+                    "max_tokens": 128000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen3-32B",
                     "tags": "LLM,CHAT,128k",
-                    "max_tokens": 8192,
+                    "max_tokens": 128000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen3-14B",
                     "tags": "LLM,CHAT,128k",
-                    "max_tokens": 8192,
+                    "max_tokens": 128000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen3-8B",
                     "tags": "LLM,CHAT,64k",
-                    "max_tokens": 8192,
+                    "max_tokens": 64000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/QVQ-72B-Preview",
                     "tags": "LLM,CHAT,IMAGE2TEXT,32k",
-                    "max_tokens": 16384,
+                    "max_tokens": 32000,
                     "model_type": "image2text",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Pro/deepseek-ai/DeepSeek-R1",
                     "tags": "LLM,CHAT,64k",
-                    "max_tokens": 16384,
+                    "max_tokens": 64000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "deepseek-ai/DeepSeek-R1",
                     "tags": "LLM,CHAT,64k",
-                    "max_tokens": 16384,
+                    "max_tokens": 64000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Pro/deepseek-ai/DeepSeek-V3",
                     "tags": "LLM,CHAT,64k",
-                    "max_tokens": 8192,
+                    "max_tokens": 64000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "deepseek-ai/DeepSeek-V3",
                     "tags": "LLM,CHAT,64k",
-                    "max_tokens": 8192,
+                    "max_tokens": 64000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Pro/deepseek-ai/DeepSeek-V3-1226",
                     "tags": "LLM,CHAT,64k",
-                    "max_tokens": 4096,
+                    "max_tokens": 64000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 16384,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-14B",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 16384,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Pro/deepseek-ai/DeepSeek-R1-Distill-Qwen-7B",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 16384,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-7B",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 16384,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Pro/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 16384,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 16384,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "deepseek-ai/DeepSeek-V2.5",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/QwQ-32B",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 32768,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen2.5-VL-72B-Instruct",
                     "tags": "LLM,CHAT,IMAGE2TEXT,128k",
-                    "max_tokens": 4096,
+                    "max_tokens": 128000,
                     "model_type": "image2text",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Pro/Qwen/Qwen2.5-VL-7B-Instruct",
                     "tags": "LLM,CHAT,IMAGE2TEXT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "image2text",
                     "is_tools": true
                 },
                 {
                     "llm_name": "THUDM/GLM-Z1-32B-0414",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "THUDM/GLM-4-32B-0414",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 8192,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "THUDM/GLM-Z1-9B-0414",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 8192,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "THUDM/GLM-4-9B-0414",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "THUDM/chatglm3-6b",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Pro/THUDM/glm-4-9b-chat",
                     "tags": "LLM,CHAT,128k",
-                    "max_tokens": 4096,
+                    "max_tokens": 128000,
                     "model_type": "chat",
                     "is_tools": false
                 },
                 {
                     "llm_name": "THUDM/GLM-Z1-Rumination-32B-0414",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": false
                 },
                 {
                     "llm_name": "THUDM/glm-4-9b-chat",
                     "tags": "LLM,CHAT,128k",
-                    "max_tokens": 4096,
+                    "max_tokens": 128000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/QwQ-32B-Preview",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 8192,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Qwen/Qwen2.5-Coder-32B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Qwen/Qwen2-VL-72B-Instruct",
                     "tags": "LLM,IMAGE2TEXT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "image2text",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Qwen/Qwen2.5-72B-Instruct-128Kt",
                     "tags": "LLM,IMAGE2TEXT,128k",
-                    "max_tokens": 4096,
+                    "max_tokens": 128000,
                     "model_type": "image2text",
                     "is_tools": false
                 },
@@ -2839,98 +2860,98 @@
                 {
                     "llm_name": "Qwen/Qwen2.5-72B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen2.5-32B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen2.5-14B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen2.5-7B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen2.5-Coder-7B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "internlm/internlm2_5-20b-chat",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "internlm/internlm2_5-7b-chat",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen2-7B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Qwen/Qwen2-1.5B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Pro/Qwen/Qwen2.5-Coder-7B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Pro/Qwen/Qwen2-VL-7B-Instruct",
                     "tags": "LLM,CHAT,IMAGE2TEXT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "image2text",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Pro/Qwen/Qwen2.5-7B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "Pro/Qwen/Qwen2-7B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": false
                 },
                 {
                     "llm_name": "Pro/Qwen/Qwen2-1.5B-Instruct",
                     "tags": "LLM,CHAT,32k",
-                    "max_tokens": 4096,
+                    "max_tokens": 32000,
                     "model_type": "chat",
                     "is_tools": false
                 },
@@ -3267,45 +3288,52 @@
             "status": "1",
             "llm": [
                 {
-                    "llm_name": "claude-opus-4-20250514",
-                    "tags": "LLM,IMAGE2TEXT,200k",
-                    "max_tokens": 204800,
-                    "model_type": "image2text",
-                    "is_tools": true
-                },
-                {
-                    "llm_name": "claude-sonnet-4-20250514",
-                    "tags": "LLM,IMAGE2TEXT,200k",
-                    "max_tokens": 204800,
-                    "model_type": "image2text",
-                    "is_tools": true
-                },
-                {
-                    "llm_name": "claude-3-7-sonnet-20250219",
-                    "tags": "LLM,IMAGE2TEXT,200k",
-                    "max_tokens": 204800,
-                    "model_type": "image2text",
-                    "is_tools": true
-                },
-                {
-                    "llm_name": "claude-3-5-sonnet-20241022",
-                    "tags": "LLM,IMAGE2TEXT,200k",
+                    "llm_name": "claude-opus-4-1-20250805",
+                    "tags": "LLM,CHAT,IMAGE2TEXT,200k",
                     "max_tokens": 204800,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
-                    "llm_name": "claude-3-opus-20240229",
-                    "tags": "LLM,IMAGE2TEXT,200k",
+                    "llm_name": "claude-opus-4-20250514",
+                    "tags": "LLM,CHAT,IMAGE2TEXT,200k",
+                    "max_tokens": 204800,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-sonnet-4-20250514",
+                    "tags": "LLM,CHAT,IMAGE2TEXT,200k",
+                    "max_tokens": 204800,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-3-7-sonnet-20250219",
+                    "tags": "LLM,CHAT,IMAGE2TEXT,200k",
+                    "max_tokens": 204800,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-3-5-sonnet-20241022",
+                    "tags": "LLM,CHAT,IMAGE2TEXT,200k",
+                    "max_tokens": 204800,
+                    "model_type": "chat",
+                    "is_tools": true
+                },
+                {
+                    "llm_name": "claude-3-5-haiku-20241022",
+                    "tags": "LLM,CHAT,IMAGE2TEXT,200k",
                     "max_tokens": 204800,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "claude-3-haiku-20240307",
-                    "tags": "LLM,IMAGE2TEXT,200k",
+                    "tags": "LLM,CHAT,IMAGE2TEXT,200k",
                     "max_tokens": 204800,
-                    "model_type": "image2text",
+                    "model_type": "chat",
                     "is_tools": true
                 }
             ]

deepdoc/parser/ppt_parser.py

@@ -87,7 +87,7 @@ class RAGFlowPptParser:
                 break
             texts = []
             for shape in sorted(
-                    slide.shapes, key=lambda x: ((x.top if x.top is not None else 0) // 10, x.left)):
+                    slide.shapes, key=lambda x: ((x.top if x.top is not None else 0) // 10, x.left if x.left is not None else 0)):
                 try:
                     txt = self.__extract(shape)
                     if txt:
@@ -96,4 +96,4 @@ class RAGFlowPptParser:
                     logging.exception(e)
             txts.append("\n".join(texts))
 
-        return txts
+        return txts

docker/migration.sh

@@ -0,0 +1,298 @@
+#!/bin/bash
+
+# RAGFlow Data Migration Script
+# Usage: ./migration.sh [backup|restore] [backup_folder]
+# 
+# This script helps you backup and restore RAGFlow Docker volumes
+# including MySQL, MinIO, Redis, and Elasticsearch data.
+
+set -e  # Exit on any error
+# Instead, we'll handle errors manually for better debugging experience
+
+# Default values
+DEFAULT_BACKUP_FOLDER="backup"
+VOLUMES=("docker_mysql_data" "docker_minio_data" "docker_redis_data" "docker_esdata01")
+BACKUP_FILES=("mysql_backup.tar.gz" "minio_backup.tar.gz" "redis_backup.tar.gz" "es_backup.tar.gz")
+
+# Function to display help information
+show_help() {
+    echo "RAGFlow Data Migration Tool"
+    echo ""
+    echo "USAGE:"
+    echo "  $0 <operation> [backup_folder]"
+    echo ""
+    echo "OPERATIONS:"
+    echo "  backup   - Create backup of all RAGFlow data volumes"
+    echo "  restore  - Restore RAGFlow data volumes from backup"
+    echo "  help     - Show this help message"
+    echo ""
+    echo "PARAMETERS:"
+    echo "  backup_folder  - Name of backup folder (default: '$DEFAULT_BACKUP_FOLDER')"
+    echo ""
+    echo "EXAMPLES:"
+    echo "  $0 backup                    # Backup to './backup' folder"
+    echo "  $0 backup my_backup          # Backup to './my_backup' folder"
+    echo "  $0 restore                   # Restore from './backup' folder"
+    echo "  $0 restore my_backup         # Restore from './my_backup' folder"
+    echo ""
+    echo "DOCKER VOLUMES:"
+    echo "  - docker_mysql_data     (MySQL database)"
+    echo "  - docker_minio_data     (MinIO object storage)"
+    echo "  - docker_redis_data     (Redis cache)"
+    echo "  - docker_esdata01       (Elasticsearch indices)"
+}
+
+# Function to check if Docker is running
+check_docker() {
+    if ! docker info >/dev/null 2>&1; then
+        echo "❌ Error: Docker is not running or not accessible"
+        echo "Please start Docker and try again"
+        exit 1
+    fi
+}
+
+# Function to check if volume exists
+volume_exists() {
+    local volume_name=$1
+    docker volume inspect "$volume_name" >/dev/null 2>&1
+}
+
+# Function to check if any containers are using the target volumes
+check_containers_using_volumes() {
+    echo "🔍 Checking for running containers that might be using target volumes..."
+    
+    # Get all running containers
+    local running_containers=$(docker ps --format "{{.Names}}")
+    
+    if [ -z "$running_containers" ]; then
+        echo "✅ No running containers found"
+        return 0
+    fi
+    
+    # Check each running container for volume usage
+    local containers_using_volumes=()
+    local volume_usage_details=()
+    
+    for container in $running_containers; do
+        # Get container's mount information
+        local mounts=$(docker inspect "$container" --format '{{range .Mounts}}{{.Source}}{{"|"}}{{end}}' 2>/dev/null || echo "")
+        
+        # Check if any of our target volumes are used by this container
+        for volume in "${VOLUMES[@]}"; do
+            if echo "$mounts" | grep -q "$volume"; then
+                containers_using_volumes+=("$container")
+                volume_usage_details+=("$container -> $volume")
+                break
+            fi
+        done
+    done
+    
+    # If any containers are using our volumes, show error and exit
+    if [ ${#containers_using_volumes[@]} -gt 0 ]; then
+        echo ""
+        echo "❌ ERROR: Found running containers using target volumes!"
+        echo ""
+        echo "📋 Running containers status:"
+        docker ps --format "table {{.Names}}\t{{.Status}}\t{{.Image}}"
+        echo ""
+        echo "🔗 Volume usage details:"
+        for detail in "${volume_usage_details[@]}"; do
+            echo "  - $detail"
+        done
+        echo ""
+        echo "🛑 SOLUTION: Stop the containers before performing backup/restore operations:"
+        echo "   docker-compose -f docker/<your-docker-compose-file>.yml down"
+        echo ""
+        echo "💡 After backup/restore, you can restart with:"
+        echo "   docker-compose -f docker/<your-docker-compose-file>.yml up -d"
+        echo ""
+        exit 1
+    fi
+    
+    echo "✅ No containers are using target volumes, safe to proceed"
+    return 0
+}
+
+# Function to confirm user action
+confirm_action() {
+    local message=$1
+    echo -n "$message (y/N): "
+    read -r response
+    case "$response" in
+        [yY]|[yY][eE][sS]) return 0 ;;
+        *) return 1 ;;
+    esac
+}
+
+# Function to perform backup
+perform_backup() {
+    local backup_folder=$1
+    
+    echo "🚀 Starting RAGFlow data backup..."
+    echo "📁 Backup folder: $backup_folder"
+    echo ""
+    
+    # Check if any containers are using the volumes
+    check_containers_using_volumes
+    
+    # Create backup folder if it doesn't exist
+    mkdir -p "$backup_folder"
+    
+    # Backup each volume
+    for i in "${!VOLUMES[@]}"; do
+        local volume="${VOLUMES[$i]}"
+        local backup_file="${BACKUP_FILES[$i]}"
+        local step=$((i + 1))
+        
+        echo "📦 Step $step/4: Backing up $volume..."
+        
+        if volume_exists "$volume"; then
+            docker run --rm \
+                -v "$volume":/source \
+                -v "$(pwd)/$backup_folder":/backup \
+                alpine tar czf "/backup/$backup_file" -C /source .
+            echo "✅ Successfully backed up $volume to $backup_folder/$backup_file"
+        else
+            echo "⚠️  Warning: Volume $volume does not exist, skipping..."
+        fi
+        echo ""
+    done
+    
+    echo "🎉 Backup completed successfully!"
+    echo "📍 Backup location: $(pwd)/$backup_folder"
+    
+    # List backup files with sizes
+    echo ""
+    echo "📋 Backup files created:"
+    for backup_file in "${BACKUP_FILES[@]}"; do
+        if [ -f "$backup_folder/$backup_file" ]; then
+            local size=$(ls -lh "$backup_folder/$backup_file" | awk '{print $5}')
+            echo "  - $backup_file ($size)"
+        fi
+    done
+}
+
+# Function to perform restore
+perform_restore() {
+    local backup_folder=$1
+    
+    echo "🔄 Starting RAGFlow data restore..."
+    echo "📁 Backup folder: $backup_folder"
+    echo ""
+    
+    # Check if any containers are using the volumes
+    check_containers_using_volumes
+    
+    # Check if backup folder exists
+    if [ ! -d "$backup_folder" ]; then
+        echo "❌ Error: Backup folder '$backup_folder' does not exist"
+        exit 1
+    fi
+    
+    # Check if all backup files exist
+    local missing_files=()
+    for backup_file in "${BACKUP_FILES[@]}"; do
+        if [ ! -f "$backup_folder/$backup_file" ]; then
+            missing_files+=("$backup_file")
+        fi
+    done
+    
+    if [ ${#missing_files[@]} -gt 0 ]; then
+        echo "❌ Error: Missing backup files:"
+        for file in "${missing_files[@]}"; do
+            echo "  - $file"
+        done
+        echo "Please ensure all backup files are present in '$backup_folder'"
+        exit 1
+    fi
+    
+    # Check for existing volumes and warn user
+    local existing_volumes=()
+    for volume in "${VOLUMES[@]}"; do
+        if volume_exists "$volume"; then
+            existing_volumes+=("$volume")
+        fi
+    done
+    
+    if [ ${#existing_volumes[@]} -gt 0 ]; then
+        echo "⚠️  WARNING: The following Docker volumes already exist:"
+        for volume in "${existing_volumes[@]}"; do
+            echo "  - $volume"
+        done
+        echo ""
+        echo "🔴 IMPORTANT: Restoring will OVERWRITE existing data!"
+        echo "💡 Recommendation: Create a backup of your current data first:"
+        echo "   $0 backup current_backup_$(date +%Y%m%d_%H%M%S)"
+        echo ""
+        
+        if ! confirm_action "Do you want to continue with the restore operation?"; then
+            echo "❌ Restore operation cancelled by user"
+            exit 0
+        fi
+    fi
+    
+    # Create volumes and restore data
+    for i in "${!VOLUMES[@]}"; do
+        local volume="${VOLUMES[$i]}"
+        local backup_file="${BACKUP_FILES[$i]}"
+        local step=$((i + 1))
+        
+        echo "🔧 Step $step/4: Restoring $volume..."
+        
+        # Create volume if it doesn't exist
+        if ! volume_exists "$volume"; then
+            echo "  📋 Creating Docker volume: $volume"
+            docker volume create "$volume"
+        else
+            echo "  📋 Using existing Docker volume: $volume"
+        fi
+        
+        # Restore data
+        echo "  📥 Restoring data from $backup_file..."
+        docker run --rm \
+            -v "$volume":/target \
+            -v "$(pwd)/$backup_folder":/backup \
+            alpine tar xzf "/backup/$backup_file" -C /target
+        
+        echo "✅ Successfully restored $volume"
+        echo ""
+    done
+    
+    echo "🎉 Restore completed successfully!"
+    echo "💡 You can now start your RAGFlow services"
+}
+
+# Main script logic
+main() {
+    # Check if Docker is available
+    check_docker
+    
+    # Parse command line arguments
+    local operation=${1:-}
+    local backup_folder=${2:-$DEFAULT_BACKUP_FOLDER}
+    
+    # Handle help or no arguments
+    if [ -z "$operation" ] || [ "$operation" = "help" ] || [ "$operation" = "-h" ] || [ "$operation" = "--help" ]; then
+        show_help
+        exit 0
+    fi
+    
+    # Validate operation
+    case "$operation" in
+        backup)
+            perform_backup "$backup_folder"
+            ;;
+        restore)
+            perform_restore "$backup_folder"
+            ;;
+        *)
+            echo "❌ Error: Invalid operation '$operation'"
+            echo ""
+            show_help
+            exit 1
+            ;;
+    esac
+}
+
+# Run main function with all arguments
+main "$@"

docs/guides/agent/agent_component_reference/agent.mdx

@@ -82,7 +82,7 @@ An integer specifying the number of previous dialogue rounds to input into the L
 This feature is used for multi-turn dialogue *only*.
 :::
 
-### Max retrieves
+### Max retries
 
 Defines the maximum number of attempts the agent will make to retry a failed task or operation before stopping or reporting failure.
 
@@ -94,6 +94,10 @@ The waiting period in seconds that the agent observes before retrying a failed t
 
 Defines the maximum number reflection rounds of the selected chat model. Defaults to 5 rounds.
 
+:::tip NOTE
+You can set the value to 1 to shorten your agent's response time.
+:::
+
 ### Output
 
 The global variable name for the output of the **Agent** component, which can be referenced by other components in the workflow.

docs/guides/migration/migrate_from_docker_compose.md

@@ -0,0 +1,108 @@
+# Data Migration Guide
+
+A common scenario is processing large datasets on a powerful instance (e.g., with a GPU) and then migrating the entire RAGFlow service to a different production environment (e.g., a CPU-only server). This guide explains how to safely back up and restore your data using our provided migration script.
+
+## Identifying Your Data
+
+By default, RAGFlow uses Docker volumes to store all persistent data, including your database, uploaded files, and search indexes. You can see these volumes by running:
+
+```bash
+docker volume ls
+```
+
+The output will look similar to this:
+
+```text
+DRIVER    VOLUME NAME
+local     docker_esdata01
+local     docker_minio_data
+local     docker_mysql_data
+local     docker_redis_data
+```
+
+These volumes contain all the data you need to migrate.
+
+## Step 1: Stop RAGFlow Services
+
+Before starting the migration, you must stop all running RAGFlow services on the **source machine**. Navigate to the project's root directory and run:
+
+```bash
+docker-compose -f docker/docker-compose.yml down
+```
+
+**Important:** Do **not** use the `-v` flag (e.g., `docker-compose down -v`), as this will delete all your data volumes. The migration script includes a check and will prevent you from running it if services are active.
+
+## Step 2: Back Up Your Data
+
+We provide a convenient script to package all your data volumes into a single backup folder.
+
+For a quick reference of the script's commands and options, you can run:
+```bash
+bash docker/migration.sh help
+```
+
+To create a backup, run the following command from the project's root directory:
+
+```bash
+bash docker/migration.sh backup
+```
+
+This will create a `backup/` folder in your project root containing compressed archives of your data volumes.
+
+You can also specify a custom name for your backup folder:
+
+```bash
+bash docker/migration.sh backup my_ragflow_backup
+```
+
+This will create a folder named `my_ragflow_backup/` instead.
+
+## Step 3: Transfer the Backup Folder
+
+Copy the entire backup folder (e.g., `backup/` or `my_ragflow_backup/`) from your source machine to the RAGFlow project directory on your **target machine**. You can use tools like `scp`, `rsync`, or a physical drive for the transfer.
+
+## Step 4: Restore Your Data
+
+On the **target machine**, ensure that RAGFlow services are not running. Then, use the migration script to restore your data from the backup folder.
+
+If your backup folder is named `backup/`, run:
+
+```bash
+bash docker/migration.sh restore
+```
+
+If you used a custom name, specify it in the command:
+
+```bash
+bash docker/migration.sh restore my_ragflow_backup
+```
+
+The script will automatically create the necessary Docker volumes and unpack the data.
+
+**Note:** If the script detects that Docker volumes with the same names already exist on the target machine, it will warn you that restoring will overwrite the existing data and ask for confirmation before proceeding.
+
+## Step 5: Start RAGFlow Services
+
+Once the restore process is complete, you can start the RAGFlow services on your new machine:
+
+```bash
+docker-compose -f docker/docker-compose.yml up -d
+```
+
+**Note:** If you already have build an service by docker-compose before, you may need to backup your data for target machine like this guide above and run like:
+
+```bash
+# Please backup by `sh docker/migration.sh backup backup_dir_name` before you do the following line.
+# !!! this line -v flag will delete the original docker volume
+docker-compose -f docker/docker-compose.yml down -v
+docker-compose -f docker/docker-compose.yml up -d
+```
+
+Your RAGFlow instance is now running with all the data from your original machine.
+
+
+
+
+
+
+

docs/references/http_api_reference.md

@@ -1118,14 +1118,14 @@ Failure:
 
 ### List documents
 
-**GET** `/api/v1/datasets/{dataset_id}/documents?page={page}&page_size={page_size}&orderby={orderby}&desc={desc}&keywords={keywords}&id={document_id}&name={document_name}`
+**GET** `/api/v1/datasets/{dataset_id}/documents?page={page}&page_size={page_size}&orderby={orderby}&desc={desc}&keywords={keywords}&id={document_id}&name={document_name}&create_time_from={timestamp}&create_time_to={timestamp}`
 
 Lists documents in a specified dataset.
 
 #### Request
 
 - Method: GET
-- URL: `/api/v1/datasets/{dataset_id}/documents?page={page}&page_size={page_size}&orderby={orderby}&desc={desc}&keywords={keywords}&id={document_id}&name={document_name}`
+- URL: `/api/v1/datasets/{dataset_id}/documents?page={page}&page_size={page_size}&orderby={orderby}&desc={desc}&keywords={keywords}&id={document_id}&name={document_name}&create_time_from={timestamp}&create_time_to={timestamp}`
 - Headers:
   - `'content-Type: application/json'`
   - `'Authorization: Bearer <YOUR_API_KEY>'`
@@ -1134,7 +1134,7 @@ Lists documents in a specified dataset.
 
 ```bash
 curl --request GET \
-     --url http://{address}/api/v1/datasets/{dataset_id}/documents?page={page}&page_size={page_size}&orderby={orderby}&desc={desc}&keywords={keywords}&id={document_id}&name={document_name} \
+     --url http://{address}/api/v1/datasets/{dataset_id}/documents?page={page}&page_size={page_size}&orderby={orderby}&desc={desc}&keywords={keywords}&id={document_id}&name={document_name}&create_time_from={timestamp}&create_time_to={timestamp} \
      --header 'Authorization: Bearer <YOUR_API_KEY>'
 ```
 
@@ -1156,6 +1156,10 @@ curl --request GET \
   Indicates whether the retrieved documents should be sorted in descending order. Defaults to `true`.
 - `id`: (*Filter parameter*), `string`  
   The ID of the document to retrieve.
+- `create_time_from`: (*Filter parameter*), `integer`
+  Unix timestamp for filtering documents created after this time. 0 means no filter. Defaults to `0`.
+- `create_time_to`: (*Filter parameter*), `integer`
+  Unix timestamp for filtering documents created before this time. 0 means no filter. Defaults to `0`.
 
 #### Response
 

docs/references/python_api_reference.md

@@ -507,7 +507,16 @@ print(doc)
 ### List documents
 
 ```python
-Dataset.list_documents(id:str =None, keywords: str=None, page: int=1, page_size:int = 30, order_by:str = "create_time", desc: bool = True) -> list[Document]
+Dataset.list_documents(
+    id: str = None,
+    keywords: str = None,
+    page: int = 1,
+    page_size: int = 30,
+    order_by: str = "create_time",
+    desc: bool = True,
+    create_time_from: int = 0,
+    create_time_to: int = 0
+) -> list[Document]
 ```
 
 Lists documents in the current dataset.
@@ -541,6 +550,12 @@ The field by which documents should be sorted. Available options:
 
 Indicates whether the retrieved documents should be sorted in descending order. Defaults to `True`.
 
+##### create_time_from: `int`
+Unix timestamp for filtering documents created after this time. 0 means no filter. Defaults to 0.
+
+##### create_time_to: `int`
+Unix timestamp for filtering documents created before this time. 0 means no filter. Defaults to 0.
+
 #### Returns
 
 - Success: A list of `Document` objects.

docs/release_notes.md

@@ -22,6 +22,39 @@ The embedding models included in a full edition are:
 These two embedding models are optimized specifically for English and Chinese, so performance may be compromised if you use them to embed documents in other languages.
 :::
 
+## v0.20.0
+
+Released on August 4, 2025.
+
+### Compatibility changes
+
+From v0.20.0 onwards, Agents are no longer compatible with earlier versions, and all existing Agents from previous versions must be rebuilt following the upgrade.
+
+### New features
+
+- Unified orchestration of both Agents and Workflows.
+- A comprehensive refactor of the Agent, greatly enhancing its capabilities and usability, with support for Multi-Agent configurations, planning and reflection, and visual functionalities.
+- Fully implemented MCP functionality, allowing for MCP Server import, Agents functioning as MCP Clients, and RAGFlow itself operating as an MCP Server.
+- Access to runtime logs for Agents.
+- Chat histories with Agents available through the management panel.
+- Integration of a new, more robust version of Infinity, enabling the auto-tagging functionality with Infinity as the underlying document engine.
+- An OpenAI-compatible API that supports file reference information.
+- Support for new models, including Kimi K2, Grok 4, and Voyage embedding.
+- RAGFlow’s codebase is now mirrored on Gitee.
+- Introduction of a new model provider, Gitee AI.
+
+### New agent templates introduced
+
+- Multi-Agent based Deep Research: Collaborative Agent teamwork led by a Lead Agent with multiple Subagents, distinct from traditional workflow orchestration.
+- An intelligent Q&A chatbot leveraging internal knowledge bases, designed for customer service and training scenarios.
+- A resume analysis template used by the RAGFlow team to screen, analyze, and record candidate information.
+- A blog generation workflow that transforms raw ideas into SEO-friendly blog content.
+- An intelligent customer service workflow.
+- A user feedback analysis template that directs user feedback to appropriate teams through semantic analysis.
+- Trip Planner: Uses web search and map MCP servers to assist with travel planning.
+- Image Lingo: Translates content from uploaded photos.
+- An information search assistant that retrieves answers from both internal knowledge bases and the web.
+
 ## v0.19.1
 
 Released on June 23, 2025.

graphrag/general/extractor.py

@@ -47,7 +47,7 @@ class Extractor:
         self._language = language
         self._entity_types = entity_types or DEFAULT_ENTITY_TYPES
 
-    @timeout(60*3)
+    @timeout(60*5)
     def _chat(self, system, history, gen_conf={}):
         hist = deepcopy(history)
         conf = deepcopy(gen_conf)

rag/app/presentation.py

@@ -42,7 +42,7 @@ class Ppt(PptParser):
                 try:
                     with BytesIO() as buffered:
                         slide.get_thumbnail(
-                            0.5, 0.5).save(
+                            0.1, 0.1).save(
                             buffered, drawing.imaging.ImageFormat.jpeg)
                         buffered.seek(0)
                         imgs.append(Image.open(buffered).copy())
@@ -135,7 +135,8 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             sections = pdf_parser(filename, binary, from_page=from_page, to_page=to_page, callback=callback)
         elif layout_recognizer == "Plain Text":
             pdf_parser = PlainParser()
-            sections, _ = pdf_parser(filename, binary, from_page=from_page, to_page=to_page, callback=callback)
+            sections, _ = pdf_parser(filename if not binary else binary, from_page=from_page, to_page=to_page,
+                                      callback=callback)
         else:
             vision_model = LLMBundle(kwargs["tenant_id"], LLMType.IMAGE2TEXT, llm_name=layout_recognizer, lang=lang)
             pdf_parser = VisionParser(vision_model=vision_model, **kwargs)

rag/llm/chat_model.py

@@ -1075,6 +1075,9 @@ class GeminiChat(Base):
         for k in list(gen_conf.keys()):
             if k not in ["temperature", "top_p", "max_tokens"]:
                 del gen_conf[k]
+            # if max_tokens exists, rename it to max_output_tokens to match Gemini's API
+            if k == "max_tokens":
+                gen_conf["max_output_tokens"] = gen_conf.pop("max_tokens")
         return gen_conf
 
     def _chat(self, history, gen_conf={}, **kwargs):

rag/llm/cv_model.py

@@ -59,6 +59,10 @@ class Base(ABC):
     def _image_prompt(self, text, images):
         if not images:
             return text
+
+        if isinstance(images, str):
+            images = [images]
+
         pmpt = [{"type": "text", "text": text}]
         for img in images:
             pmpt.append({
@@ -518,6 +522,7 @@ class GeminiCV(Base):
     def chat_streamly(self, system, history, gen_conf, images=[]):
         from transformers import GenerationConfig
         ans = ""
+        response = None
         try:
             response = self.model.generate_content(
                 self._form_history(system, history, images),
@@ -533,8 +538,11 @@ class GeminiCV(Base):
         except Exception as e:
             yield ans + "\n**ERROR**: " + str(e)
 
-        yield response._chunks[-1].usage_metadata.total_token_count
-
+        if response and hasattr(response, "usage_metadata") and hasattr(response.usage_metadata, "total_token_count"):
+            yield response.usage_metadata.total_token_count
+        else:
+            yield 0
+            
 
 class NvidiaCV(Base):
     _FACTORY_NAME = "NVIDIA"
@@ -616,15 +624,18 @@ class NvidiaCV(Base):
             return "**ERROR**: " + str(e), 0
 
     def chat_streamly(self, system, history, gen_conf, images=[], **kwargs):
+        total_tokens = 0
         try:
             response = self._request(self._form_history(system, history, images), gen_conf)
             cnt = response["choices"][0]["message"]["content"]
+            if "usage" in response and "total_tokens" in response["usage"]:
+                total_tokens += response["usage"]["total_tokens"]
             for resp in cnt:
                 yield resp
         except Exception as e:
             yield "\n**ERROR**: " + str(e)
 
-        yield response["usage"]["total_tokens"]
+        yield total_tokens
 
 
 class AnthropicCV(Base):
@@ -795,4 +806,4 @@ class GoogleCV(AnthropicCV, GeminiCV):
                 yield ans
         else:
             for ans in GeminiCV.chat_streamly(self, system, history, gen_conf, images):
-                yield ans
+                yield ans

rag/nlp/__init__.py

@@ -634,6 +634,16 @@ def concat_img(img1, img2):
         return img2
     if not img1 and not img2:
         return None
+    
+    if img1 is img2:
+        return img1
+    
+    if isinstance(img1, Image.Image) and isinstance(img2, Image.Image):
+        pixel_data1 = img1.tobytes()
+        pixel_data2 = img2.tobytes()
+        if pixel_data1 == pixel_data2:
+            return img1
+
     width1, height1 = img1.size
     width2, height2 = img2.size
 
@@ -643,7 +653,6 @@ def concat_img(img1, img2):
 
     new_image.paste(img1, (0, 0))
     new_image.paste(img2, (0, height1))
-
     return new_image
 
 

rag/svr/task_executor.py

@@ -284,7 +284,7 @@ async def build_chunks(task, progress_callback):
         try:
             d = copy.deepcopy(document)
             d.update(chunk)
-            d["id"] = xxhash.xxh64((chunk["content_with_weight"] + str(d["doc_id"])).encode("utf-8")).hexdigest()
+            d["id"] = xxhash.xxh64((chunk["content_with_weight"] + str(d["doc_id"])).encode("utf-8", "surrogatepass")).hexdigest()
             d["create_time"] = str(datetime.now()).replace("T", " ")[:19]
             d["create_timestamp_flt"] = datetime.now().timestamp()
             if not d.get("image"):
@@ -420,7 +420,6 @@ def init_kb(row, vector_size: int):
     return settings.docStoreConn.createIdx(idxnm, row.get("kb_id", ""), vector_size)
 
 
-@timeout(60*20)
 async def embedding(docs, mdl, parser_config=None, callback=None):
     if parser_config is None:
         parser_config = {}
@@ -441,10 +440,15 @@ async def embedding(docs, mdl, parser_config=None, callback=None):
         tts = np.concatenate([vts for _ in range(len(tts))], axis=0)
         tk_count += c
 
+    @timeout(5)
+    def batch_encode(txts):
+        nonlocal mdl
+        return mdl.encode([truncate(c, mdl.max_length-10) for c in txts])
+
     cnts_ = np.array([])
     for i in range(0, len(cnts), EMBEDDING_BATCH_SIZE):
         async with embed_limiter:
-            vts, c = await trio.to_thread.run_sync(lambda: mdl.encode([truncate(c, mdl.max_length-10) for c in cnts[i: i + EMBEDDING_BATCH_SIZE]]))
+            vts, c = await trio.to_thread.run_sync(lambda: batch_encode(cnts[i: i + EMBEDDING_BATCH_SIZE]))
         if len(cnts_) == 0:
             cnts_ = vts
         else:

rag/utils/redis_conn.py

@@ -227,9 +227,20 @@ class RedisDB:
         """https://redis.io/docs/latest/commands/xreadgroup/"""
         for _ in range(3):
             try:
-                group_info = self.REDIS.xinfo_groups(queue_name)
-                if not any(gi["name"] == group_name for gi in group_info):
-                    self.REDIS.xgroup_create(queue_name, group_name, id="0", mkstream=True)
+
+                try:
+                    group_info = self.REDIS.xinfo_groups(queue_name)
+                    if not any(gi["name"] == group_name for gi in group_info):
+                        self.REDIS.xgroup_create(queue_name, group_name, id="0", mkstream=True)
+                except redis.exceptions.ResponseError as e:
+                    if "no such key" in str(e).lower():
+                        self.REDIS.xgroup_create(queue_name, group_name, id="0", mkstream=True)
+                    elif "busygroup" in str(e).lower():
+                        logging.warning("Group already exists, continue.")
+                        pass
+                    else:
+                        raise
+
                 args = {
                     "groupname": group_name,
                     "consumername": consumer_name,
@@ -338,8 +349,8 @@ class RedisDB:
             logging.warning("RedisDB.delete " + str(key) + " got exception: " + str(e))
             self.__open__()
         return False
-    
-    
+
+
 REDIS_CONN = RedisDB()
 
 

rag/utils/s3_conn.py

@@ -123,7 +123,7 @@ class RAGFlowS3:
 
     @use_prefix_path
     @use_default_bucket
-    def put(self, bucket, fnm, binary):
+    def put(self, bucket, fnm, binary, **kwargs):
         logging.debug(f"bucket name {bucket}; filename :{fnm}:")
         for _ in range(1):
             try:
@@ -140,7 +140,7 @@ class RAGFlowS3:
 
     @use_prefix_path
     @use_default_bucket
-    def rm(self, bucket, fnm):
+    def rm(self, bucket, fnm, **kwargs):
         try:
             self.conn.delete_object(Bucket=bucket, Key=fnm)
         except Exception:
@@ -148,7 +148,7 @@ class RAGFlowS3:
 
     @use_prefix_path
     @use_default_bucket
-    def get(self, bucket, fnm):
+    def get(self, bucket, fnm, **kwargs):
         for _ in range(1):
             try:
                 r = self.conn.get_object(Bucket=bucket, Key=fnm)
@@ -162,7 +162,7 @@ class RAGFlowS3:
 
     @use_prefix_path
     @use_default_bucket
-    def obj_exist(self, bucket, fnm):
+    def obj_exist(self, bucket, fnm, **kwargs):
         try:
             if self.conn.head_object(Bucket=bucket, Key=fnm):
                 return True
@@ -174,7 +174,7 @@ class RAGFlowS3:
 
     @use_prefix_path
     @use_default_bucket
-    def get_presigned_url(self, bucket, fnm, expires):
+    def get_presigned_url(self, bucket, fnm, expires, **kwargs):
         for _ in range(10):
             try:
                 r = self.conn.generate_presigned_url('get_object',

sdk/python/ragflow_sdk/modules/dataset.py

@@ -63,8 +63,30 @@ class DataSet(Base):
             return doc_list
         raise Exception(res.get("message"))
 
-    def list_documents(self, id: str | None = None, name: str | None = None, keywords: str | None = None, page: int = 1, page_size: int = 30, orderby: str = "create_time", desc: bool = True):
-        res = self.get(f"/datasets/{self.id}/documents", params={"id": id, "name": name, "keywords": keywords, "page": page, "page_size": page_size, "orderby": orderby, "desc": desc})
+    def list_documents(
+        self,
+        id: str | None = None,
+        name: str | None = None,
+        keywords: str | None = None,
+        page: int = 1,
+        page_size: int = 30,
+        orderby: str = "create_time",
+        desc: bool = True,
+        create_time_from: int = 0,
+        create_time_to: int = 0,
+    ):
+        params = {
+            "id": id,
+            "name": name,
+            "keywords": keywords,
+            "page": page,
+            "page_size": page_size,
+            "orderby": orderby,
+            "desc": desc,
+            "create_time_from": create_time_from,
+            "create_time_to": create_time_to,
+        }
+        res = self.get(f"/datasets/{self.id}/documents", params=params)
         res = res.json()
         documents = []
         if res.get("code") == 0:

web/src/components/knowledge-base-item.tsx

@@ -1,9 +1,13 @@
 import { DocumentParserType } from '@/constants/knowledge';
 import { useTranslate } from '@/hooks/common-hooks';
 import { useFetchKnowledgeList } from '@/hooks/knowledge-hooks';
+import { useBuildQueryVariableOptions } from '@/pages/agent/hooks/use-get-begin-query';
 import { UserOutlined } from '@ant-design/icons';
 import { Avatar as AntAvatar, Form, Select, Space } from 'antd';
+import { toLower } from 'lodash';
+import { useMemo } from 'react';
 import { useFormContext } from 'react-hook-form';
+import { useTranslation } from 'react-i18next';
 import { RAGFlowAvatar } from './ragflow-avatar';
 import { FormControl, FormField, FormItem, FormLabel } from './ui/form';
 import { MultiSelect } from './ui/multi-select';
@@ -66,9 +70,13 @@ const KnowledgeBaseItem = ({
 
 export default KnowledgeBaseItem;
 
-export function KnowledgeBaseFormField() {
+export function KnowledgeBaseFormField({
+  showVariable = false,
+}: {
+  showVariable?: boolean;
+}) {
   const form = useFormContext();
-  const { t } = useTranslate('chat');
+  const { t } = useTranslation();
 
   const { list: knowledgeList } = useFetchKnowledgeList(true);
 
@@ -76,6 +84,8 @@ export function KnowledgeBaseFormField() {
     (x) => x.parser_id !== DocumentParserType.Tag,
   );
 
+  const nextOptions = useBuildQueryVariableOptions();
+
   const knowledgeOptions = filteredKnowledgeList.map((x) => ({
     label: x.name,
     value: x.id,
@@ -84,18 +94,48 @@ export function KnowledgeBaseFormField() {
     ),
   }));
 
+  const options = useMemo(() => {
+    if (showVariable) {
+      return [
+        {
+          label: t('knowledgeDetails.dataset'),
+          options: knowledgeOptions,
+        },
+        ...nextOptions.map((x) => {
+          return {
+            ...x,
+            options: x.options
+              .filter((y) => toLower(y.type).includes('string'))
+              .map((x) => ({
+                ...x,
+                icon: () => (
+                  <RAGFlowAvatar
+                    className="size-4 mr-2"
+                    avatar={x.label}
+                    name={x.label}
+                  />
+                ),
+              })),
+          };
+        }),
+      ];
+    }
+
+    return knowledgeOptions;
+  }, [knowledgeOptions, nextOptions, showVariable, t]);
+
   return (
     <FormField
       control={form.control}
       name="kb_ids"
       render={({ field }) => (
         <FormItem>
-          <FormLabel>{t('knowledgeBases')}</FormLabel>
+          <FormLabel>{t('chat.knowledgeBases')}</FormLabel>
           <FormControl>
             <MultiSelect
-              options={knowledgeOptions}
+              options={options}
               onValueChange={field.onChange}
-              placeholder={t('knowledgeBasesMessage')}
+              placeholder={t('chat.knowledgeBasesMessage')}
               variant="inverted"
               maxCount={100}
               defaultValue={field.value}

web/src/components/originui/number-input.tsx

@@ -63,7 +63,7 @@ const NumberInput: React.FC<NumberInputProps> = ({
     >
       <button
         type="button"
-        className="w-10 p-2 text-white focus:outline-none border-r-[1px]"
+        className="w-10 p-2 focus:outline-none border-r-[1px]"
         onClick={handleDecrement}
         style={style}
       >
@@ -74,12 +74,12 @@ const NumberInput: React.FC<NumberInputProps> = ({
         value={value}
         onInput={handleInput}
         onChange={handleChange}
-        className="w-full flex-1 text-center bg-transparent text-white focus:outline-none"
+        className="w-full flex-1 text-center bg-transparent focus:outline-none"
         style={style}
       />
       <button
         type="button"
-        className="w-10 p-2 text-white focus:outline-none border-l-[1px]"
+        className="w-10 p-2 focus:outline-none border-l-[1px]"
         onClick={handleIncrement}
         style={style}
       >

web/src/components/ui/accordion.tsx

@@ -28,8 +28,11 @@ function AccordionItem({
 function AccordionTrigger({
   className,
   children,
+  hideDownIcon = false,
   ...props
-}: React.ComponentProps<typeof AccordionPrimitive.Trigger>) {
+}: React.ComponentProps<typeof AccordionPrimitive.Trigger> & {
+  hideDownIcon?: boolean;
+}) {
   return (
     <AccordionPrimitive.Header className="flex">
       <AccordionPrimitive.Trigger
@@ -41,7 +44,9 @@ function AccordionTrigger({
         {...props}
       >
         {children}
-        <ChevronDownIcon className="text-muted-foreground pointer-events-none size-4 shrink-0 translate-y-0.5 transition-transform duration-200" />
+        {!hideDownIcon && (
+          <ChevronDownIcon className="text-muted-foreground pointer-events-none size-4 shrink-0 translate-y-0.5 transition-transform duration-200" />
+        )}
       </AccordionPrimitive.Trigger>
     </AccordionPrimitive.Header>
   );

web/src/components/ui/multi-select.tsx

@@ -1,3 +1,4 @@
+// https://github.com/sersavan/shadcn-multi-select-component
 // src/components/multi-select.tsx
 
 import { cva, type VariantProps } from 'class-variance-authority';
@@ -29,6 +30,51 @@ import {
 import { Separator } from '@/components/ui/separator';
 import { cn } from '@/lib/utils';
 
+export type MultiSelectOptionType = {
+  label: React.ReactNode;
+  value: string;
+  disabled?: boolean;
+  icon?: React.ComponentType<{ className?: string }>;
+};
+
+export type MultiSelectGroupOptionType = {
+  label: React.ReactNode;
+  options: MultiSelectOptionType[];
+};
+
+function MultiCommandItem({
+  option,
+  isSelected,
+  toggleOption,
+}: {
+  option: MultiSelectOptionType;
+  isSelected: boolean;
+  toggleOption(value: string): void;
+}) {
+  return (
+    <CommandItem
+      key={option.value}
+      onSelect={() => toggleOption(option.value)}
+      className="cursor-pointer"
+    >
+      <div
+        className={cn(
+          'mr-2 flex h-4 w-4 items-center justify-center rounded-sm border border-primary',
+          isSelected
+            ? 'bg-primary text-primary-foreground'
+            : 'opacity-50 [&_svg]:invisible',
+        )}
+      >
+        <CheckIcon className="h-4 w-4" />
+      </div>
+      {option.icon && (
+        <option.icon className="mr-2 h-4 w-4 text-muted-foreground" />
+      )}
+      <span>{option.label}</span>
+    </CommandItem>
+  );
+}
+
 /**
  * Variants for the multi-select component to handle different styles.
  * Uses class-variance-authority (cva) to define different styles based on "variant" prop.
@@ -63,14 +109,7 @@ interface MultiSelectProps
    * An array of option objects to be displayed in the multi-select component.
    * Each option object has a label, value, and an optional icon.
    */
-  options: {
-    /** The text to display for the option. */
-    label: string;
-    /** The unique value associated with the option. */
-    value: string;
-    /** Optional icon component to display alongside the option. */
-    icon?: React.ComponentType<{ className?: string }>;
-  }[];
+  options: (MultiSelectGroupOptionType | MultiSelectOptionType)[];
 
   /**
    * Callback function triggered when the selected values change.
@@ -144,6 +183,11 @@ export const MultiSelect = React.forwardRef<
     const [isPopoverOpen, setIsPopoverOpen] = React.useState(false);
     const [isAnimating, setIsAnimating] = React.useState(false);
 
+    const flatOptions = React.useMemo(() => {
+      return options.flatMap((option) =>
+        'options' in option ? option.options : [option],
+      );
+    }, [options]);
     const handleInputKeyDown = (
       event: React.KeyboardEvent<HTMLInputElement>,
     ) => {
@@ -181,10 +225,10 @@ export const MultiSelect = React.forwardRef<
     };
 
     const toggleAll = () => {
-      if (selectedValues.length === options.length) {
+      if (selectedValues.length === flatOptions.length) {
         handleClear();
       } else {
-        const allValues = options.map((option) => option.value);
+        const allValues = flatOptions.map((option) => option.value);
         setSelectedValues(allValues);
         onValueChange(allValues);
       }
@@ -210,7 +254,7 @@ export const MultiSelect = React.forwardRef<
               <div className="flex justify-between items-center w-full">
                 <div className="flex flex-wrap items-center">
                   {selectedValues?.slice(0, maxCount)?.map((value) => {
-                    const option = options.find((o) => o.value === value);
+                    const option = flatOptions.find((o) => o.value === value);
                     const IconComponent = option?.icon;
                     return (
                       <Badge
@@ -304,7 +348,7 @@ export const MultiSelect = React.forwardRef<
                   <div
                     className={cn(
                       'mr-2 flex h-4 w-4 items-center justify-center rounded-sm border border-primary',
-                      selectedValues.length === options.length
+                      selectedValues.length === flatOptions.length
                         ? 'bg-primary text-primary-foreground'
                         : 'opacity-50 [&_svg]:invisible',
                     )}
@@ -313,32 +357,38 @@ export const MultiSelect = React.forwardRef<
                   </div>
                   <span>(Select All)</span>
                 </CommandItem>
-                {options.map((option) => {
-                  const isSelected = selectedValues.includes(option.value);
-                  return (
-                    <CommandItem
-                      key={option.value}
-                      onSelect={() => toggleOption(option.value)}
-                      className="cursor-pointer"
-                    >
-                      <div
-                        className={cn(
-                          'mr-2 flex h-4 w-4 items-center justify-center rounded-sm border border-primary',
-                          isSelected
-                            ? 'bg-primary text-primary-foreground'
-                            : 'opacity-50 [&_svg]:invisible',
-                        )}
-                      >
-                        <CheckIcon className="h-4 w-4" />
-                      </div>
-                      {option.icon && (
-                        <option.icon className="mr-2 h-4 w-4 text-muted-foreground" />
-                      )}
-                      <span>{option.label}</span>
-                    </CommandItem>
-                  );
-                })}
+                {!options.some((x) => 'options' in x) &&
+                  (options as unknown as MultiSelectOptionType[]).map(
+                    (option) => {
+                      const isSelected = selectedValues.includes(option.value);
+                      return (
+                        <MultiCommandItem
+                          option={option}
+                          key={option.value}
+                          isSelected={isSelected}
+                          toggleOption={toggleOption}
+                        ></MultiCommandItem>
+                      );
+                    },
+                  )}
               </CommandGroup>
+              {options.every((x) => 'options' in x) &&
+                options.map((x, idx) => (
+                  <CommandGroup heading={x.label} key={idx}>
+                    {x.options.map((option) => {
+                      const isSelected = selectedValues.includes(option.value);
+
+                      return (
+                        <MultiCommandItem
+                          option={option}
+                          key={option.value}
+                          isSelected={isSelected}
+                          toggleOption={toggleOption}
+                        ></MultiCommandItem>
+                      );
+                    })}
+                  </CommandGroup>
+                ))}
               <CommandSeparator />
               <CommandGroup>
                 <div className="flex items-center justify-between">

web/src/constants/common.ts

@@ -49,8 +49,8 @@ export const LanguageList = [
   'Japanese',
   'Portuguese BR',
   'German',
+  'French',
 ];
-
 export const LanguageMap = {
   English: 'English',
   Chinese: '简体中文',
@@ -61,6 +61,7 @@ export const LanguageMap = {
   Japanese: '日本語',
   'Portuguese BR': 'Português BR',
   German: 'German',
+  French: 'Français',
 };
 
 export enum LanguageAbbreviation {
@@ -73,6 +74,7 @@ export enum LanguageAbbreviation {
   Vi = 'vi',
   PtBr = 'pt-BR',
   De = 'de',
+  Fr = 'fr',
 }
 
 export const LanguageAbbreviationMap = {
@@ -85,6 +87,7 @@ export const LanguageAbbreviationMap = {
   [LanguageAbbreviation.Ja]: '日本語',
   [LanguageAbbreviation.PtBr]: 'Português BR',
   [LanguageAbbreviation.De]: 'Deutsch',
+  [LanguageAbbreviation.Fr]: 'Français',
 };
 
 export const LanguageTranslationMap = {
@@ -97,6 +100,7 @@ export const LanguageTranslationMap = {
   Japanese: 'ja',
   'Portuguese BR': 'pt-br',
   German: 'de',
+  French: 'fr',
 };
 
 export enum FileMimeType {

web/src/hooks/logic-hooks.ts

@@ -353,7 +353,12 @@ export const useHandleMessageInputChange = () => {
 export const useSelectDerivedMessages = () => {
   const [derivedMessages, setDerivedMessages] = useState<IMessage[]>([]);
 
-  const ref = useScrollToBottom(derivedMessages);
+  const messageContainerRef = useRef<HTMLDivElement>(null);
+
+  const { scrollRef, scrollToBottom } = useScrollToBottom(
+    derivedMessages,
+    messageContainerRef,
+  );
 
   const addNewestQuestion = useCallback(
     (message: Message, answer: string = '') => {
@@ -492,7 +497,8 @@ export const useSelectDerivedMessages = () => {
   }, [setDerivedMessages]);
 
   return {
-    ref,
+    scrollRef,
+    messageContainerRef,
     derivedMessages,
     setDerivedMessages,
     addNewestQuestion,
@@ -503,6 +509,7 @@ export const useSelectDerivedMessages = () => {
     addNewestOneAnswer,
     removeMessagesAfterCurrentMessage,
     removeAllMessages,
+    scrollToBottom,
   };
 };
 

web/src/hooks/logic-hooks/navigate-hooks.ts

@@ -35,9 +35,12 @@ export const useNavigatePage = () => {
     navigate(Routes.Chats);
   }, [navigate]);
 
-  const navigateToChat = useCallback(() => {
-    navigate(Routes.Chat);
-  }, [navigate]);
+  const navigateToChat = useCallback(
+    (id: string) => () => {
+      navigate(`${Routes.Chat}/${id}`);
+    },
+    [navigate],
+  );
 
   const navigateToAgents = useCallback(() => {
     navigate(Routes.Agents);

web/src/hooks/use-chat-request.ts

@@ -1,9 +1,23 @@
+import message from '@/components/ui/message';
 import { ChatSearchParams } from '@/constants/chat';
 import { IDialog } from '@/interfaces/database/chat';
-import chatService from '@/services/chat-service';
-import { useQuery } from '@tanstack/react-query';
+import chatService from '@/services/next-chat-service ';
+import { useMutation, useQuery, useQueryClient } from '@tanstack/react-query';
+import { useDebounce } from 'ahooks';
 import { useCallback, useMemo } from 'react';
-import { history, useSearchParams } from 'umi';
+import { useTranslation } from 'react-i18next';
+import { useParams, useSearchParams } from 'umi';
+import {
+  useGetPaginationWithRouter,
+  useHandleSearchChange,
+} from './logic-hooks';
+
+export const enum ChatApiAction {
+  FetchDialogList = 'fetchDialogList',
+  RemoveDialog = 'removeDialog',
+  SetDialog = 'setDialog',
+  FetchDialog = 'fetchDialog',
+}
 
 export const useGetChatSearchParams = () => {
   const [currentQueryParameters] = useSearchParams();
@@ -39,37 +53,128 @@ export const useClickDialogCard = () => {
   return { handleClickDialog };
 };
 
-export const useFetchDialogList = (pureFetch = false) => {
-  const { handleClickDialog } = useClickDialogCard();
-  const { dialogId } = useGetChatSearchParams();
+export const useFetchDialogList = () => {
+  const { searchString, handleInputChange } = useHandleSearchChange();
+  const { pagination, setPagination } = useGetPaginationWithRouter();
+  const debouncedSearchString = useDebounce(searchString, { wait: 500 });
 
   const {
     data,
     isFetching: loading,
     refetch,
-  } = useQuery<IDialog[]>({
-    queryKey: ['fetchDialogList'],
-    initialData: [],
+  } = useQuery<{ dialogs: IDialog[]; total: number }>({
+    queryKey: [
+      ChatApiAction.FetchDialogList,
+      {
+        debouncedSearchString,
+        ...pagination,
+      },
+    ],
+    initialData: { dialogs: [], total: 0 },
     gcTime: 0,
     refetchOnWindowFocus: false,
-    queryFn: async (...params) => {
-      console.log('🚀 ~ queryFn: ~ params:', params);
-      const { data } = await chatService.listDialog();
+    queryFn: async () => {
+      const { data } = await chatService.listDialog({
+        keywords: debouncedSearchString,
+        page_size: pagination.pageSize,
+        page: pagination.current,
+      });
 
+      return data?.data ?? { dialogs: [], total: 0 };
+    },
+  });
+
+  const onInputChange: React.ChangeEventHandler<HTMLInputElement> = useCallback(
+    (e) => {
+      handleInputChange(e);
+    },
+    [handleInputChange],
+  );
+
+  return {
+    data,
+    loading,
+    refetch,
+    searchString,
+    handleInputChange: onInputChange,
+    pagination: { ...pagination, total: data?.total },
+    setPagination,
+  };
+};
+
+export const useRemoveDialog = () => {
+  const queryClient = useQueryClient();
+  const { t } = useTranslation();
+
+  const {
+    data,
+    isPending: loading,
+    mutateAsync,
+  } = useMutation({
+    mutationKey: [ChatApiAction.RemoveDialog],
+    mutationFn: async (dialogIds: string[]) => {
+      const { data } = await chatService.removeDialog({ dialogIds });
       if (data.code === 0) {
-        const list: IDialog[] = data.data;
-        if (!pureFetch) {
-          if (list.length > 0) {
-            if (list.every((x) => x.id !== dialogId)) {
-              handleClickDialog(data.data[0].id);
-            }
-          } else {
-            history.push('/chat');
-          }
-        }
-      }
+        queryClient.invalidateQueries({ queryKey: ['fetchDialogList'] });
 
-      return data?.data ?? [];
+        message.success(t('message.deleted'));
+      }
+      return data.code;
+    },
+  });
+
+  return { data, loading, removeDialog: mutateAsync };
+};
+
+export const useSetDialog = () => {
+  const queryClient = useQueryClient();
+  const { t } = useTranslation();
+
+  const {
+    data,
+    isPending: loading,
+    mutateAsync,
+  } = useMutation({
+    mutationKey: [ChatApiAction.SetDialog],
+    mutationFn: async (params: Partial<IDialog>) => {
+      const { data } = await chatService.setDialog(params);
+      if (data.code === 0) {
+        queryClient.invalidateQueries({
+          exact: false,
+          queryKey: [ChatApiAction.FetchDialogList],
+        });
+
+        message.success(
+          t(`message.${params.dialog_id ? 'modified' : 'created'}`),
+        );
+      }
+      return data?.code;
+    },
+  });
+
+  return { data, loading, setDialog: mutateAsync };
+};
+
+export const useFetchDialog = () => {
+  const { id } = useParams();
+
+  const {
+    data,
+    isFetching: loading,
+    refetch,
+  } = useQuery<IDialog>({
+    queryKey: [ChatApiAction.FetchDialog, id],
+    gcTime: 0,
+    initialData: {} as IDialog,
+    enabled: !!id,
+    refetchOnWindowFocus: false,
+    queryFn: async () => {
+      const { data } = await chatService.getDialog(
+        { params: { dialogId: id } },
+        true,
+      );
+
+      return data?.data ?? ({} as IDialog);
     },
   });
 

web/src/locales/config.ts

@@ -6,6 +6,7 @@ import { LanguageAbbreviation } from '@/constants/common';
 import translation_de from './de';
 import translation_en from './en';
 import translation_es from './es';
+import translation_fr from './fr';
 import translation_id from './id';
 import translation_ja from './ja';
 import translation_pt_br from './pt-br';
@@ -24,6 +25,7 @@ const resources = {
   [LanguageAbbreviation.Vi]: translation_vi,
   [LanguageAbbreviation.PtBr]: translation_pt_br,
   [LanguageAbbreviation.De]: translation_de,
+  [LanguageAbbreviation.Fr]: translation_fr,
 };
 const enFlattened = flattenObject(translation_en);
 const viFlattened = flattenObject(translation_vi);
@@ -33,6 +35,7 @@ const jaFlattened = flattenObject(translation_ja);
 const pt_brFlattened = flattenObject(translation_pt_br);
 const zh_traditionalFlattened = flattenObject(translation_zh_traditional);
 const deFlattened = flattenObject(translation_de);
+const frFlattened = flattenObject(translation_fr);
 export const translationTable = createTranslationTable(
   [
     enFlattened,
@@ -43,6 +46,7 @@ export const translationTable = createTranslationTable(
     jaFlattened,
     pt_brFlattened,
     deFlattened,
+    frFlattened,
   ],
   [
     'English',
@@ -53,6 +57,7 @@ export const translationTable = createTranslationTable(
     'ja',
     'pt-BR',
     'Deutsch',
+    'French',
   ],
 );
 i18n

web/src/locales/en.ts

@@ -39,6 +39,13 @@ export default {
       nextPage: 'Next',
       add: 'Add',
       promptPlaceholder: `Please input or use / to quickly insert variables.`,
+      mcp: {
+        namePlaceholder: 'My MCP Server',
+        nameRequired:
+          'It must be 1–64 characters long and can only contain letters, numbers, hyphens, and underscores.',
+        urlPlaceholder: 'https://api.example.com/v1/mcp',
+        tokenPlaceholder: 'e.g. eyJhbGciOiJIUzI1Ni...',
+      },
     },
     login: {
       login: 'Sign in',
@@ -555,6 +562,7 @@ This auto-tagging feature enhances retrieval by adding another layer of domain-s
       tavilyApiKeyHelp: 'How to get it?',
       crossLanguage: 'Cross-language search',
       crossLanguageTip: `Select one or more languages for cross‑language search. If no language is selected, the system searches with the original query.`,
+      createChat: 'Create chat',
     },
     setting: {
       profile: 'Profile',
@@ -1322,6 +1330,7 @@ This delimiter is used to split the input text into several text pieces echo of
       logTimeline: {
         begin: 'Ready to begin',
         agent: 'Agent is thinking',
+        userFillUp: 'Waiting for you',
         retrieval: 'Looking up knowledge',
         message: 'Agent says',
         awaitResponse: 'Waiting for you',

web/src/locales/zh.ts

@@ -843,7 +843,7 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
       relevant: '是否相关',
       rewriteQuestion: '问题优化',
       begin: '开始',
-      message: '静态消息',
+      message: '回复消息',
       blank: '空',
       createFromNothing: '从无到有',
       addItem: '新增',
@@ -1245,7 +1245,7 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
       modeTip: '模式定义了工作流的启动方式。',
       beginInputTip: '通过定义输入参数，此内容可以被后续流程中的其他组件访问。',
       query: '查询变量',
-      agent: 'Agent',
+      agent: '智能体',
       agentDescription: '构建具备推理、工具调用和多智能体协同的智能体组件。',
       maxRecords: '最大记录数',
       createAgent: 'Create Agent',
@@ -1265,6 +1265,7 @@ General：实体和关系提取提示来自 GitHub - microsoft/graphrag：基于
       subject: '主题',
       logTimeline: {
         begin: '准备开始',
+        userFillUp: '等你输入',
         agent: '智能体正在思考',
         retrieval: '查找知识',
         message: '回复',

web/src/pages/agent/canvas/node/dropdown/next-step-dropdown.tsx

@@ -11,6 +11,11 @@ import {
   DropdownMenuLabel,
   DropdownMenuTrigger,
 } from '@/components/ui/dropdown-menu';
+import {
+  Tooltip,
+  TooltipContent,
+  TooltipTrigger,
+} from '@/components/ui/tooltip';
 import { IModalProps } from '@/interfaces/common';
 import { Operator } from '@/pages/agent/constant';
 import { AgentInstanceContext, HandleContext } from '@/pages/agent/context';
@@ -33,19 +38,26 @@ function OperatorItemList({ operators }: OperatorItemProps) {
     <ul className="space-y-2">
       {operators.map((x) => {
         return (
-          <DropdownMenuItem
-            key={x}
-            className="hover:bg-background-card py-1 px-3 cursor-pointer rounded-sm flex gap-2 items-center justify-start"
-            onClick={addCanvasNode(x, {
-              nodeId,
-              id,
-              position,
-            })}
-            onSelect={() => hideModal?.()}
-          >
-            <OperatorIcon name={x}></OperatorIcon>
-            {t(`flow.${lowerFirst(x)}`)}
-          </DropdownMenuItem>
+          <Tooltip key={x}>
+            <TooltipTrigger asChild>
+              <DropdownMenuItem
+                key={x}
+                className="hover:bg-background-card py-1 px-3 cursor-pointer rounded-sm flex gap-2 items-center justify-start"
+                onClick={addCanvasNode(x, {
+                  nodeId,
+                  id,
+                  position,
+                })}
+                onSelect={() => hideModal?.()}
+              >
+                <OperatorIcon name={x}></OperatorIcon>
+                {t(`flow.${lowerFirst(x)}`)}
+              </DropdownMenuItem>
+            </TooltipTrigger>
+            <TooltipContent side="right">
+              <p>{t(`flow.${lowerFirst(x)}Description`)}</p>
+            </TooltipContent>
+          </Tooltip>
         );
       })}
     </ul>

web/src/pages/agent/canvas/node/retrieval-node.tsx

@@ -2,11 +2,11 @@ import { RAGFlowAvatar } from '@/components/ragflow-avatar';
 import { useFetchKnowledgeList } from '@/hooks/knowledge-hooks';
 import { IRetrievalNode } from '@/interfaces/database/flow';
 import { NodeProps, Position } from '@xyflow/react';
-import { Flex } from 'antd';
 import classNames from 'classnames';
 import { get } from 'lodash';
 import { memo, useMemo } from 'react';
 import { NodeHandleId } from '../../constant';
+import { useGetVariableLabelByValue } from '../../hooks/use-get-begin-query';
 import { CommonHandle } from './handle';
 import { LeftHandleStyle, RightHandleStyle } from './handle-icon';
 import styles from './index.less';
@@ -21,6 +21,7 @@ function InnerRetrievalNode({
   selected,
 }: NodeProps<IRetrievalNode>) {
   const knowledgeBaseIds: string[] = get(data, 'form.kb_ids', []);
+  console.log('🚀 ~ InnerRetrievalNode ~ knowledgeBaseIds:', knowledgeBaseIds);
   const { list: knowledgeList } = useFetchKnowledgeList(true);
   const knowledgeBases = useMemo(() => {
     return knowledgeBaseIds.map((x) => {
@@ -33,6 +34,8 @@ function InnerRetrievalNode({
     });
   }, [knowledgeList, knowledgeBaseIds]);
 
+  const getLabel = useGetVariableLabelByValue(id);
+
   return (
     <ToolBar selected={selected} id={id} label={data.label}>
       <NodeWrapper selected={selected}>
@@ -63,25 +66,27 @@ function InnerRetrievalNode({
             [styles.nodeHeader]: knowledgeBaseIds.length > 0,
           })}
         ></NodeHeader>
-        <Flex vertical gap={8}>
-          {knowledgeBases.map((knowledge) => {
+        <section className="flex flex-col gap-2">
+          {knowledgeBaseIds.map((id) => {
+            const item = knowledgeList.find((y) => id === y.id);
+            const label = getLabel(id);
+
             return (
-              <div className={styles.nodeText} key={knowledge.id}>
-                <Flex align={'center'} gap={6}>
+              <div className={styles.nodeText} key={id}>
+                <div className="flex items-center gap-1.5">
                   <RAGFlowAvatar
                     className="size-6 rounded-lg"
-                    avatar={knowledge.avatar}
-                    name={knowledge.name || 'CN'}
+                    avatar={id}
+                    name={item?.name || (label as string) || 'CN'}
                     isPerson={true}
                   />
-                  <Flex className={styles.knowledgeNodeName} flex={1}>
-                    {knowledge.name}
-                  </Flex>
-                </Flex>
+
+                  <div className={'truncate flex-1'}>{label || item?.name}</div>
+                </div>
               </div>
             );
           })}
-        </Flex>
+        </section>
       </NodeWrapper>
     </ToolBar>
   );

web/src/pages/agent/chat/box.tsx

@@ -13,19 +13,17 @@ import {
   useUploadCanvasFileWithProgress,
 } from '@/hooks/use-agent-request';
 import { useFetchUserInfo } from '@/hooks/user-setting-hooks';
-import { Message } from '@/interfaces/database/chat';
 import { buildMessageUuidWithRole } from '@/utils/chat';
-import { get } from 'lodash';
-import { memo, useCallback, useMemo } from 'react';
+import { memo, useCallback } from 'react';
 import { useParams } from 'umi';
 import DebugContent from '../debug-content';
-import { BeginQuery } from '../interface';
-import { buildBeginQueryWithObject } from '../utils';
+import { useAwaitCompentData } from '../hooks/use-chat-logic';
 
 function AgentChatBox() {
   const {
     value,
-    ref,
+    scrollRef,
+    messageContainerRef,
     sendLoading,
     derivedMessages,
     handleInputChange,
@@ -43,33 +41,12 @@ function AgentChatBox() {
   const { data: canvasInfo } = useFetchAgent();
   const { id: canvasId } = useParams();
   const { uploadCanvasFile, loading } = useUploadCanvasFileWithProgress();
-  const getInputs = useCallback((message: Message) => {
-    return get(message, 'data.inputs', {}) as Record<string, BeginQuery>;
-  }, []);
 
-  const buildInputList = useCallback(
-    (message: Message) => {
-      return Object.entries(getInputs(message)).map(([key, val]) => {
-        return {
-          ...val,
-          key,
-        };
-      });
-    },
-    [getInputs],
-  );
-
-  const handleOk = useCallback(
-    (message: Message) => (values: BeginQuery[]) => {
-      const inputs = getInputs(message);
-      const nextInputs = buildBeginQueryWithObject(inputs, values);
-      sendFormMessage({
-        inputs: nextInputs,
-        id: canvasId,
-      });
-    },
-    [canvasId, getInputs, sendFormMessage],
-  );
+  const { buildInputList, handleOk, isWaitting } = useAwaitCompentData({
+    derivedMessages,
+    sendFormMessage,
+    canvasId: canvasId as string,
+  });
 
   const handleUploadFile: NonNullable<FileUploadProps['onUpload']> =
     useCallback(
@@ -79,20 +56,11 @@ function AgentChatBox() {
       },
       [appendUploadResponseList, uploadCanvasFile],
     );
-  const isWaitting = useMemo(() => {
-    const temp = derivedMessages?.some((message, i) => {
-      const flag =
-        message.role === MessageType.Assistant &&
-        derivedMessages.length - 1 === i &&
-        message.data;
-      return flag;
-    });
-    return temp;
-  }, [derivedMessages]);
+
   return (
     <>
       <section className="flex flex-1 flex-col px-5 h-[90vh]">
-        <div className="flex-1 overflow-auto">
+        <div className="flex-1 overflow-auto" ref={messageContainerRef}>
           <div>
             {/* <Spin spinning={sendLoading}> */}
             {derivedMessages?.map((message, i) => {
@@ -139,7 +107,7 @@ function AgentChatBox() {
             })}
             {/* </Spin> */}
           </div>
-          <div ref={ref.scrollRef} />
+          <div ref={scrollRef} />
         </div>
         <NextMessageInput
           value={value}

web/src/pages/agent/chat/use-send-agent-message.ts

@@ -198,12 +198,14 @@ export const useSendAgentMessage = (
   const prologue = useGetBeginNodePrologue();
   const {
     derivedMessages,
-    ref,
+    scrollRef,
+    messageContainerRef,
     removeLatestMessage,
     removeMessageById,
     addNewestOneQuestion,
     addNewestOneAnswer,
     removeAllMessages,
+    scrollToBottom,
   } = useSelectDerivedMessages();
   const { addEventList: addEventListFun } = useContext(AgentChatLogContext);
   const {
@@ -269,7 +271,7 @@ export const useSendAgentMessage = (
 
   const sendFormMessage = useCallback(
     (body: { id?: string; inputs: Record<string, BeginQuery> }) => {
-      send(body);
+      send({ ...body, session_id: sessionId });
       addNewestOneQuestion({
         content: Object.entries(body.inputs)
           .map(([key, val]) => `${key}: ${val.value}`)
@@ -277,7 +279,7 @@ export const useSendAgentMessage = (
         role: MessageType.User,
       });
     },
-    [addNewestOneQuestion, send],
+    [addNewestOneQuestion, send, sessionId],
   );
 
   // reset session
@@ -303,7 +305,18 @@ export const useSendAgentMessage = (
       });
     }
     addNewestOneQuestion({ ...msgBody, files: fileList });
-  }, [value, done, addNewestOneQuestion, fileList, setValue, sendMessage]);
+    setTimeout(() => {
+      scrollToBottom();
+    }, 100);
+  }, [
+    value,
+    done,
+    addNewestOneQuestion,
+    fileList,
+    setValue,
+    sendMessage,
+    scrollToBottom,
+  ]);
 
   useEffect(() => {
     const { content, id } = findMessageFromList(answerList);
@@ -337,7 +350,8 @@ export const useSendAgentMessage = (
     value,
     sendLoading: !done,
     derivedMessages,
-    ref,
+    scrollRef,
+    messageContainerRef,
     handlePressEnter,
     handleInputChange,
     removeMessageById,

web/src/pages/agent/constant.tsx

@@ -48,8 +48,6 @@ export const BeginId = 'begin';
 export enum Operator {
   Begin = 'Begin',
   Retrieval = 'Retrieval',
-  Generate = 'Generate',
-  Answer = 'Answer',
   Categorize = 'Categorize',
   Message = 'Message',
   Relevant = 'Relevant',
@@ -78,7 +76,6 @@ export enum Operator {
   Note = 'Note',
   Crawler = 'Crawler',
   Invoke = 'Invoke',
-  Template = 'Template',
   Email = 'Email',
   Iteration = 'Iteration',
   IterationStart = 'IterationItem',
@@ -100,15 +97,12 @@ export const CommonOperatorList = Object.values(Operator).filter(
 
 export const AgentOperatorList = [
   Operator.Retrieval,
-  Operator.Generate,
-  Operator.Answer,
   Operator.Categorize,
   Operator.Message,
   Operator.RewriteQuestion,
   Operator.KeywordExtract,
   Operator.Switch,
   Operator.Concentrator,
-  Operator.Template,
   Operator.Iteration,
   Operator.WaitingDialogue,
   Operator.Note,
@@ -119,12 +113,6 @@ export const componentMenuList = [
   {
     name: Operator.Retrieval,
   },
-  {
-    name: Operator.Generate,
-  },
-  {
-    name: Operator.Answer,
-  },
   {
     name: Operator.Categorize,
   },
@@ -144,9 +132,6 @@ export const componentMenuList = [
   {
     name: Operator.Concentrator,
   },
-  {
-    name: Operator.Template,
-  },
   {
     name: Operator.Iteration,
   },
@@ -660,7 +645,7 @@ export const initialAgentValues = {
   max_retries: 3,
   delay_after_error: 1,
   visual_files_var: '',
-  max_rounds: 5,
+  max_rounds: 1,
   exception_method: '',
   exception_goto: [],
   exception_default_value: '',
@@ -796,19 +781,16 @@ export const CategorizeAnchorPointPositions = [
 // no connection lines are allowed between key and value
 export const RestrictedUpstreamMap = {
   [Operator.Begin]: [Operator.Relevant],
-  [Operator.Categorize]: [Operator.Begin, Operator.Categorize, Operator.Answer],
-  [Operator.Answer]: [Operator.Begin, Operator.Answer, Operator.Message],
+  [Operator.Categorize]: [Operator.Begin, Operator.Categorize],
   [Operator.Retrieval]: [Operator.Begin, Operator.Retrieval],
-  [Operator.Generate]: [Operator.Begin, Operator.Relevant],
   [Operator.Message]: [
     Operator.Begin,
     Operator.Message,
-    Operator.Generate,
     Operator.Retrieval,
     Operator.RewriteQuestion,
     Operator.Categorize,
   ],
-  [Operator.Relevant]: [Operator.Begin, Operator.Answer],
+  [Operator.Relevant]: [Operator.Begin],
   [Operator.RewriteQuestion]: [
     Operator.Begin,
     Operator.Message,
@@ -843,7 +825,6 @@ export const RestrictedUpstreamMap = {
   [Operator.Crawler]: [Operator.Begin],
   [Operator.Note]: [],
   [Operator.Invoke]: [Operator.Begin],
-  [Operator.Template]: [Operator.Begin, Operator.Relevant],
   [Operator.Email]: [Operator.Begin],
   [Operator.Iteration]: [Operator.Begin],
   [Operator.IterationStart]: [Operator.Begin],
@@ -861,8 +842,6 @@ export const NodeMap = {
   [Operator.Begin]: 'beginNode',
   [Operator.Categorize]: 'categorizeNode',
   [Operator.Retrieval]: 'retrievalNode',
-  [Operator.Generate]: 'generateNode',
-  [Operator.Answer]: 'logicNode',
   [Operator.Message]: 'messageNode',
   [Operator.Relevant]: 'relevantNode',
   [Operator.RewriteQuestion]: 'rewriteNode',
@@ -890,7 +869,6 @@ export const NodeMap = {
   [Operator.Note]: 'noteNode',
   [Operator.Crawler]: 'ragNode',
   [Operator.Invoke]: 'ragNode',
-  [Operator.Template]: 'templateNode',
   [Operator.Email]: 'ragNode',
   [Operator.Iteration]: 'group',
   [Operator.IterationStart]: 'iterationStartNode',
@@ -924,9 +902,7 @@ export const BeginQueryTypeIconMap = {
 
 export const NoDebugOperatorsList = [
   Operator.Begin,
-  Operator.Answer,
   Operator.Concentrator,
-  Operator.Template,
   Operator.Message,
   Operator.RewriteQuestion,
   Operator.Switch,

web/src/pages/agent/form-sheet/form-config-map.tsx

@@ -1,8 +1,6 @@
-import { z } from 'zod';
 import { Operator } from '../constant';
 import AgentForm from '../form/agent-form';
 import AkShareForm from '../form/akshare-form';
-import AnswerForm from '../form/answer-form';
 import ArXivForm from '../form/arxiv-form';
 import BaiduFanyiForm from '../form/baidu-fanyi-form';
 import BaiduForm from '../form/baidu-form';
@@ -15,7 +13,6 @@ import DeepLForm from '../form/deepl-form';
 import DuckDuckGoForm from '../form/duckduckgo-form';
 import EmailForm from '../form/email-form';
 import ExeSQLForm from '../form/exesql-form';
-import GenerateForm from '../form/generate-form';
 import GithubForm from '../form/github-form';
 import GoogleForm from '../form/google-form';
 import GoogleScholarForm from '../form/google-scholar-form';
@@ -34,7 +31,6 @@ import StringTransformForm from '../form/string-transform-form';
 import SwitchForm from '../form/switch-form';
 import TavilyExtractForm from '../form/tavily-extract-form';
 import TavilyForm from '../form/tavily-form';
-import TemplateForm from '../form/template-form';
 import ToolForm from '../form/tool-form';
 import TuShareForm from '../form/tushare-form';
 import UserFillUpForm from '../form/user-fill-up-form';
@@ -49,12 +45,6 @@ export const FormConfigMap = {
   [Operator.Retrieval]: {
     component: RetrievalForm,
   },
-  [Operator.Generate]: {
-    component: GenerateForm,
-  },
-  [Operator.Answer]: {
-    component: AnswerForm,
-  },
   [Operator.Categorize]: {
     component: CategorizeForm,
   },
@@ -75,8 +65,6 @@ export const FormConfigMap = {
   },
   [Operator.Agent]: {
     component: AgentForm,
-    defaultValues: {},
-    schema: z.object({}),
   },
   [Operator.Baidu]: {
     component: BaiduForm,
@@ -107,8 +95,6 @@ export const FormConfigMap = {
   },
   [Operator.DeepL]: {
     component: DeepLForm,
-    defaultValues: {},
-    schema: z.object({}),
   },
   [Operator.GitHub]: {
     component: GithubForm,
@@ -152,9 +138,6 @@ export const FormConfigMap = {
   [Operator.Note]: {
     component: () => <></>,
   },
-  [Operator.Template]: {
-    component: TemplateForm,
-  },
   [Operator.Email]: {
     component: EmailForm,
   },

web/src/pages/agent/form/answer-form/index.tsx

@@ -1,5 +0,0 @@
-const AnswerForm = () => {
-  return <div></div>;
-};
-
-export default AnswerForm;

web/src/pages/agent/form/categorize-form/index.tsx

@@ -8,6 +8,7 @@ import { useForm } from 'react-hook-form';
 import { initialCategorizeValues } from '../../constant';
 import { INextOperatorForm } from '../../interface';
 import { buildOutputList } from '../../utils/build-output-list';
+import { FormWrapper } from '../components/form-wrapper';
 import { Output } from '../components/output';
 import { QueryVariable } from '../components/query-variable';
 import DynamicCategorize from './dynamic-categorize';
@@ -31,12 +32,7 @@ function CategorizeForm({ node }: INextOperatorForm) {
 
   return (
     <Form {...form}>
-      <form
-        className="space-y-6 p-5 "
-        onSubmit={(e) => {
-          e.preventDefault();
-        }}
-      >
+      <FormWrapper>
         <FormContainer>
           <QueryVariable></QueryVariable>
           <LargeModelFormField></LargeModelFormField>
@@ -44,7 +40,7 @@ function CategorizeForm({ node }: INextOperatorForm) {
         <MessageHistoryWindowSizeFormField></MessageHistoryWindowSizeFormField>
         <DynamicCategorize nodeId={node?.id}></DynamicCategorize>
         <Output list={outputList}></Output>
-      </form>
+      </FormWrapper>
     </Form>
   );
 }

web/src/pages/agent/form/code-form/index.tsx

@@ -2,6 +2,7 @@ import Editor, { loader } from '@monaco-editor/react';
 import { INextOperatorForm } from '../../interface';
 
 import { FormContainer } from '@/components/form-container';
+import { useIsDarkTheme } from '@/components/theme-provider';
 import {
   Form,
   FormControl,
@@ -46,6 +47,7 @@ function CodeForm({ node }: INextOperatorForm) {
   const formData = node?.data.form as ICodeForm;
   const { t } = useTranslation();
   const values = useValues(node);
+  const isDarkTheme = useIsDarkTheme();
 
   const form = useForm<FormSchemaType>({
     defaultValues: values,
@@ -94,7 +96,7 @@ function CodeForm({ node }: INextOperatorForm) {
               <FormControl>
                 <Editor
                   height={300}
-                  theme="vs-dark"
+                  theme={isDarkTheme ? 'vs-dark' : 'vs'}
                   language={formData.lang}
                   options={{
                     minimap: { enabled: false },

web/src/pages/agent/form/components/prompt-editor/index.tsx

@@ -26,6 +26,7 @@ import { useLexicalComposerContext } from '@lexical/react/LexicalComposerContext
 import { Variable } from 'lucide-react';
 import { ReactNode, useCallback, useState } from 'react';
 import { useTranslation } from 'react-i18next';
+import { PasteHandlerPlugin } from './paste-handler-plugin';
 import theme from './theme';
 import { VariableNode } from './variable-node';
 import { VariableOnChangePlugin } from './variable-on-change-plugin';
@@ -103,9 +104,12 @@ function PromptContent({
         </div>
       )}
       <ContentEditable
-        className={cn('relative px-2 py-1 focus-visible:outline-none', {
-          'min-h-40': multiLine,
-        })}
+        className={cn(
+          'relative px-2 py-1 focus-visible:outline-none max-h-[50vh] overflow-auto',
+          {
+            'min-h-40': multiLine,
+          },
+        )}
         onBlur={handleBlur}
         onFocus={handleFocus}
       />
@@ -169,6 +173,7 @@ export function PromptEditor({
           ErrorBoundary={LexicalErrorBoundary}
         />
         <VariablePickerMenuPlugin value={value}></VariablePickerMenuPlugin>
+        <PasteHandlerPlugin />
         <VariableOnChangePlugin
           onChange={onValueChange}
         ></VariableOnChangePlugin>

web/src/pages/agent/form/components/prompt-editor/paste-handler-plugin.tsx

@@ -0,0 +1,83 @@
+import { useLexicalComposerContext } from '@lexical/react/LexicalComposerContext';
+import {
+  $createParagraphNode,
+  $createTextNode,
+  $getSelection,
+  $isRangeSelection,
+  PASTE_COMMAND,
+} from 'lexical';
+import { useEffect } from 'react';
+
+function PasteHandlerPlugin() {
+  const [editor] = useLexicalComposerContext();
+  useEffect(() => {
+    const removeListener = editor.registerCommand(
+      PASTE_COMMAND,
+      (clipboardEvent: ClipboardEvent) => {
+        const clipboardData = clipboardEvent.clipboardData;
+        if (!clipboardData) {
+          return false;
+        }
+
+        const text = clipboardData.getData('text/plain');
+        if (!text) {
+          return false;
+        }
+
+        // Check if text contains line breaks
+        if (text.includes('\n')) {
+          editor.update(() => {
+            const selection = $getSelection();
+            if (selection && $isRangeSelection(selection)) {
+              // Normalize line breaks, merge multiple consecutive line breaks into a single line break
+              const normalizedText = text.replace(/\n{2,}/g, '\n');
+
+              // Clear current selection
+              selection.removeText();
+
+              // Create a paragraph node to contain all content
+              const paragraph = $createParagraphNode();
+
+              // Split text by line breaks
+              const lines = normalizedText.split('\n');
+
+              // Process each line
+              lines.forEach((lineText, index) => {
+                // Add line text (if any)
+                if (lineText) {
+                  const textNode = $createTextNode(lineText);
+                  paragraph.append(textNode);
+                }
+
+                // If not the last line, add a line break
+                if (index < lines.length - 1) {
+                  const lineBreak = $createTextNode('\n');
+                  paragraph.append(lineBreak);
+                }
+              });
+
+              // Insert paragraph
+              selection.insertNodes([paragraph]);
+            }
+          });
+
+          // Prevent default paste behavior
+          clipboardEvent.preventDefault();
+          return true;
+        }
+
+        // If no line breaks, use default behavior
+        return false;
+      },
+      4,
+    );
+
+    return () => {
+      removeListener();
+    };
+  }, [editor]);
+
+  return null;
+}
+
+export { PasteHandlerPlugin };

web/src/pages/agent/form/concentrator-form/index.tsx

@@ -1,17 +0,0 @@
-import { Form } from 'antd';
-import { IOperatorForm } from '../../interface';
-
-const ConcentratorForm = ({ onValuesChange, form }: IOperatorForm) => {
-  return (
-    <Form
-      name="basic"
-      labelCol={{ span: 8 }}
-      wrapperCol={{ span: 16 }}
-      autoComplete="off"
-      form={form}
-      onValuesChange={onValuesChange}
-    ></Form>
-  );
-};
-
-export default ConcentratorForm;

web/src/pages/agent/form/generate-form/index.tsx

@@ -1,78 +0,0 @@
-import { NextLLMSelect } from '@/components/llm-select/next';
-import { MessageHistoryWindowSizeFormField } from '@/components/message-history-window-size-item';
-import {
-  Form,
-  FormControl,
-  FormField,
-  FormItem,
-  FormLabel,
-  FormMessage,
-} from '@/components/ui/form';
-import { Switch } from '@/components/ui/switch';
-import { useTranslation } from 'react-i18next';
-import { INextOperatorForm } from '../../interface';
-import { PromptEditor } from '../components/prompt-editor';
-
-const GenerateForm = ({ form }: INextOperatorForm) => {
-  const { t } = useTranslation();
-
-  return (
-    <Form {...form}>
-      <form
-        className="space-y-6"
-        onSubmit={(e) => {
-          e.preventDefault();
-        }}
-      >
-        <FormField
-          control={form.control}
-          name="llm_id"
-          render={({ field }) => (
-            <FormItem>
-              <FormLabel tooltip={t('chat.modelTip')}>
-                {t('chat.model')}
-              </FormLabel>
-              <FormControl>
-                <NextLLMSelect {...field} />
-              </FormControl>
-              <FormMessage />
-            </FormItem>
-          )}
-        />
-        <FormField
-          control={form.control}
-          name="prompt"
-          render={({ field }) => (
-            <FormItem>
-              <FormLabel tooltip={t('flow.promptTip')}>
-                {t('flow.systemPrompt')}
-              </FormLabel>
-              <FormControl>
-                <PromptEditor {...field} />
-              </FormControl>
-              <FormMessage />
-            </FormItem>
-          )}
-        />
-        <FormField
-          control={form.control}
-          name="cite"
-          render={({ field }) => (
-            <FormItem>
-              <FormLabel tooltip={t('flow.citeTip')}>
-                {t('flow.cite')}
-              </FormLabel>
-              <FormControl>
-                <Switch {...field} />
-              </FormControl>
-              <FormMessage />
-            </FormItem>
-          )}
-        />
-        <MessageHistoryWindowSizeFormField></MessageHistoryWindowSizeFormField>
-      </form>
-    </Form>
-  );
-};
-
-export default GenerateForm;

web/src/pages/agent/form/invoke-form/index.less

@@ -1,44 +0,0 @@
-.editableRow {
-  :global(.editable-cell) {
-    position: relative;
-  }
-
-  :global(.editable-cell-value-wrap) {
-    padding: 5px 12px;
-    cursor: pointer;
-    height: 30px !important;
-  }
-  &:hover {
-    :global(.editable-cell-value-wrap) {
-      padding: 4px 11px;
-      border: 1px solid #d9d9d9;
-      border-radius: 2px;
-    }
-  }
-}
-
-.dynamicParameterVariable {
-  background-color: #ebe9e950;
-  :global(.ant-collapse-content) {
-    background-color: #f6f6f634;
-  }
-  :global(.ant-collapse-content-box) {
-    padding: 0 !important;
-  }
-  margin-bottom: 20px;
-  .title {
-    font-weight: 600;
-    font-size: 16px;
-  }
-  .variableType {
-    width: 30%;
-  }
-  .variableValue {
-    flex: 1;
-  }
-
-  .addButton {
-    color: rgb(22, 119, 255);
-    font-weight: 600;
-  }
-}

web/src/pages/agent/form/iteration-form/index.tsx

@@ -6,6 +6,7 @@ import { useForm, useWatch } from 'react-hook-form';
 import { z } from 'zod';
 import { VariableType } from '../../constant';
 import { INextOperatorForm } from '../../interface';
+import { FormWrapper } from '../components/form-wrapper';
 import { Output } from '../components/output';
 import { QueryVariable } from '../components/query-variable';
 import { DynamicOutput } from './dynamic-output';
@@ -39,12 +40,7 @@ function IterationForm({ node }: INextOperatorForm) {
 
   return (
     <Form {...form}>
-      <form
-        className="space-y-6 p-4"
-        onSubmit={(e) => {
-          e.preventDefault();
-        }}
-      >
+      <FormWrapper>
         <FormContainer>
           <QueryVariable
             name="items_ref"
@@ -53,7 +49,7 @@ function IterationForm({ node }: INextOperatorForm) {
         </FormContainer>
         <DynamicOutput node={node}></DynamicOutput>
         <Output list={outputList}></Output>
-      </form>
+      </FormWrapper>
     </Form>
   );
 }

web/src/pages/agent/form/message-form/index.tsx

@@ -15,6 +15,7 @@ import { useFieldArray, useForm } from 'react-hook-form';
 import { useTranslation } from 'react-i18next';
 import { z } from 'zod';
 import { INextOperatorForm } from '../../interface';
+import { FormWrapper } from '../components/form-wrapper';
 import { PromptEditor } from '../components/prompt-editor';
 import { useValues } from './use-values';
 import { useWatchFormChange } from './use-watch-change';
@@ -48,13 +49,7 @@ function MessageForm({ node }: INextOperatorForm) {
 
   return (
     <Form {...form}>
-      <form
-        className="space-y-5 px-5 "
-        autoComplete="off"
-        onSubmit={(e) => {
-          e.preventDefault();
-        }}
-      >
+      <FormWrapper>
         <FormContainer>
           <FormItem>
             <FormLabel tooltip={t('flow.msgTip')}>{t('flow.msg')}</FormLabel>
@@ -98,7 +93,7 @@ function MessageForm({ node }: INextOperatorForm) {
             <FormMessage />
           </FormItem>
         </FormContainer>
-      </form>
+      </FormWrapper>
     </Form>
   );
 }

web/src/pages/agent/form/retrieval-form/next.tsx

@@ -97,7 +97,7 @@ function RetrievalForm({ node }: INextOperatorForm) {
       <FormWrapper>
         <FormContainer>
           <QueryVariable></QueryVariable>
-          <KnowledgeBaseFormField></KnowledgeBaseFormField>
+          <KnowledgeBaseFormField showVariable></KnowledgeBaseFormField>
         </FormContainer>
         <Collapse title={<div>Advanced Settings</div>}>
           <FormContainer>

web/src/pages/agent/form/string-transform-form/index.tsx

@@ -20,6 +20,7 @@ import {
   initialStringTransformValues,
 } from '../../constant';
 import { INextOperatorForm } from '../../interface';
+import { FormWrapper } from '../components/form-wrapper';
 import { Output, transferOutputs } from '../components/output';
 import { PromptEditor } from '../components/prompt-editor';
 import { QueryVariable } from '../components/query-variable';
@@ -76,13 +77,7 @@ function StringTransformForm({ node }: INextOperatorForm) {
 
   return (
     <Form {...form}>
-      <form
-        className="space-y-5 px-5 "
-        autoComplete="off"
-        onSubmit={(e) => {
-          e.preventDefault();
-        }}
-      >
+      <FormWrapper>
         <FormContainer>
           <FormField
             control={form.control}
@@ -157,7 +152,7 @@ function StringTransformForm({ node }: INextOperatorForm) {
             render={() => <div></div>}
           />
         </FormContainer>
-      </form>
+      </FormWrapper>
       <div className="p-5">
         <Output list={outputList}></Output>
       </div>

web/src/pages/agent/form/switch-form/index.tsx

@@ -28,6 +28,7 @@ import {
 } from '../../constant';
 import { useBuildQueryVariableOptions } from '../../hooks/use-get-begin-query';
 import { IOperatorForm } from '../../interface';
+import { FormWrapper } from '../components/form-wrapper';
 import { useValues } from './use-values';
 import { useWatchFormChange } from './use-watch-change';
 
@@ -249,12 +250,7 @@ function SwitchForm({ node }: IOperatorForm) {
 
   return (
     <Form {...form}>
-      <form
-        className="space-y-6 p-5 "
-        onSubmit={(e) => {
-          e.preventDefault();
-        }}
-      >
+      <FormWrapper>
         {fields.map((field, index) => {
           const name = `${ConditionKey}.${index}`;
           const conditions: Array<any> = form.getValues(`${name}.${ItemKey}`);
@@ -323,7 +319,7 @@ function SwitchForm({ node }: IOperatorForm) {
         >
           Add
         </BlockButton>
-      </form>
+      </FormWrapper>
     </Form>
   );
 }

web/src/pages/agent/form/template-form/index.tsx

@@ -1,24 +0,0 @@
-import { Form } from 'antd';
-import { useTranslation } from 'react-i18next';
-import { IOperatorForm } from '../../interface';
-import { PromptEditor } from '../components/prompt-editor';
-
-const TemplateForm = ({ onValuesChange, form }: IOperatorForm) => {
-  const { t } = useTranslation();
-
-  return (
-    <Form
-      name="basic"
-      autoComplete="off"
-      form={form}
-      onValuesChange={onValuesChange}
-      layout={'vertical'}
-    >
-      <Form.Item name={['content']} label={t('flow.content')}>
-        <PromptEditor></PromptEditor>
-      </Form.Item>
-    </Form>
-  );
-};
-
-export default TemplateForm;

web/src/pages/agent/form/tool-form/retrieval-form/index.tsx

@@ -11,6 +11,7 @@ import { zodResolver } from '@hookform/resolvers/zod';
 import { useForm } from 'react-hook-form';
 import { z } from 'zod';
 import { DescriptionField } from '../../components/description-field';
+import { FormWrapper } from '../../components/form-wrapper';
 import {
   EmptyResponseField,
   RetrievalPartialSchema,
@@ -35,15 +36,10 @@ const RetrievalForm = () => {
 
   return (
     <Form {...form}>
-      <form
-        className="space-y-6 p-4"
-        onSubmit={(e) => {
-          e.preventDefault();
-        }}
-      >
+      <FormWrapper>
         <FormContainer>
           <DescriptionField></DescriptionField>
-          <KnowledgeBaseFormField></KnowledgeBaseFormField>
+          <KnowledgeBaseFormField showVariable></KnowledgeBaseFormField>
         </FormContainer>
         <Collapse title={<div>Advanced Settings</div>}>
           <FormContainer>
@@ -58,7 +54,7 @@ const RetrievalForm = () => {
             <UseKnowledgeGraphFormField name="use_kg"></UseKnowledgeGraphFormField>
           </FormContainer>
         </Collapse>
-      </form>
+      </FormWrapper>
     </Form>
   );
 };

web/src/pages/agent/hooks.tsx

@@ -1,6 +1,7 @@
 import {
   Connection,
   Edge,
+  getOutgoers,
   Node,
   Position,
   ReactFlowInstance,
@@ -15,9 +16,6 @@ import { get, lowerFirst, omit } from 'lodash';
 import { UseFormReturn } from 'react-hook-form';
 import { useTranslation } from 'react-i18next';
 import {
-  NodeMap,
-  Operator,
-  RestrictedUpstreamMap,
   initialAgentValues,
   initialAkShareValues,
   initialArXivValues,
@@ -33,7 +31,6 @@ import {
   initialDuckValues,
   initialEmailValues,
   initialExeSqlValues,
-  initialGenerateValues,
   initialGithubValues,
   initialGoogleScholarValues,
   initialGoogleValues,
@@ -48,15 +45,19 @@ import {
   initialRelevantValues,
   initialRetrievalValues,
   initialRewriteQuestionValues,
+  initialStringTransformValues,
   initialSwitchValues,
   initialTavilyExtractValues,
   initialTavilyValues,
-  initialTemplateValues,
   initialTuShareValues,
+  initialUserFillUpValues,
   initialWaitingDialogueValues,
   initialWenCaiValues,
   initialWikipediaValues,
   initialYahooFinanceValues,
+  NodeMap,
+  Operator,
+  RestrictedUpstreamMap,
 } from './constant';
 import useGraphStore, { RFState } from './store';
 import {
@@ -92,8 +93,6 @@ export const useInitializeOperatorParams = () => {
     return {
       [Operator.Begin]: initialBeginValues,
       [Operator.Retrieval]: initialRetrievalValues,
-      [Operator.Generate]: { ...initialGenerateValues, llm_id: llmId },
-      [Operator.Answer]: {},
       [Operator.Categorize]: { ...initialCategorizeValues, llm_id: llmId },
       [Operator.Relevant]: { ...initialRelevantValues, llm_id: llmId },
       [Operator.RewriteQuestion]: {
@@ -128,7 +127,6 @@ export const useInitializeOperatorParams = () => {
       [Operator.Note]: initialNoteValues,
       [Operator.Crawler]: initialCrawlerValues,
       [Operator.Invoke]: initialInvokeValues,
-      [Operator.Template]: initialTemplateValues,
       [Operator.Email]: initialEmailValues,
       [Operator.Iteration]: initialIterationValues,
       [Operator.IterationStart]: initialIterationValues,
@@ -137,6 +135,9 @@ export const useInitializeOperatorParams = () => {
       [Operator.Agent]: { ...initialAgentValues, llm_id: llmId },
       [Operator.TavilySearch]: initialTavilyValues,
       [Operator.TavilyExtract]: initialTavilyExtractValues,
+      [Operator.Tool]: {},
+      [Operator.UserFillUp]: initialUserFillUpValues,
+      [Operator.StringTransform]: initialStringTransformValues,
     };
   }, [llmId]);
 
@@ -333,9 +334,8 @@ export const useHandleFormValuesChange = (
 };
 
 export const useValidateConnection = () => {
-  const { getOperatorTypeFromId, getParentIdById } = useGraphStore(
-    (state) => state,
-  );
+  const { getOperatorTypeFromId, getParentIdById, edges, nodes } =
+    useGraphStore((state) => state);
 
   const isSameNodeChild = useCallback(
     (connection: Connection | Edge) => {
@@ -349,6 +349,27 @@ export const useValidateConnection = () => {
     [getParentIdById],
   );
 
+  const hasCanvasCycle = useCallback(
+    (connection: Connection | Edge) => {
+      const target = nodes.find((node) => node.id === connection.target);
+      const hasCycle = (node: RAGFlowNodeType, visited = new Set()) => {
+        if (visited.has(node.id)) return false;
+
+        visited.add(node.id);
+
+        for (const outgoer of getOutgoers(node, nodes, edges)) {
+          if (outgoer.id === connection.source) return true;
+          if (hasCycle(outgoer, visited)) return true;
+        }
+      };
+
+      if (target?.id === connection.source) return false;
+
+      return target ? !hasCycle(target) : false;
+    },
+    [edges, nodes],
+  );
+
   // restricted lines cannot be connected successfully.
   const isValidConnection = useCallback(
     (connection: Connection | Edge) => {
@@ -365,10 +386,11 @@ export const useValidateConnection = () => {
         RestrictedUpstreamMap[
           getOperatorTypeFromId(connection.source) as Operator
         ]?.every((x) => x !== getOperatorTypeFromId(connection.target)) &&
-        isSameNodeChild(connection);
+        isSameNodeChild(connection) &&
+        hasCanvasCycle(connection);
       return ret;
     },
-    [getOperatorTypeFromId, isSameNodeChild],
+    [getOperatorTypeFromId, hasCanvasCycle, isSameNodeChild],
   );
 
   return isValidConnection;

web/src/pages/agent/hooks/use-add-node.ts

@@ -23,7 +23,6 @@ import {
   initialDuckValues,
   initialEmailValues,
   initialExeSqlValues,
-  initialGenerateValues,
   initialGithubValues,
   initialGoogleScholarValues,
   initialGoogleValues,
@@ -43,7 +42,6 @@ import {
   initialSwitchValues,
   initialTavilyExtractValues,
   initialTavilyValues,
-  initialTemplateValues,
   initialTuShareValues,
   initialUserFillUpValues,
   initialWaitingDialogueValues,
@@ -70,8 +68,6 @@ export const useInitializeOperatorParams = () => {
     return {
       [Operator.Begin]: initialBeginValues,
       [Operator.Retrieval]: initialRetrievalValues,
-      [Operator.Generate]: { ...initialGenerateValues, llm_id: llmId },
-      [Operator.Answer]: {},
       [Operator.Categorize]: { ...initialCategorizeValues, llm_id: llmId },
       [Operator.Relevant]: { ...initialRelevantValues, llm_id: llmId },
       [Operator.RewriteQuestion]: {
@@ -106,7 +102,6 @@ export const useInitializeOperatorParams = () => {
       [Operator.Note]: initialNoteValues,
       [Operator.Crawler]: initialCrawlerValues,
       [Operator.Invoke]: initialInvokeValues,
-      [Operator.Template]: initialTemplateValues,
       [Operator.Email]: initialEmailValues,
       [Operator.Iteration]: initialIterationValues,
       [Operator.IterationStart]: initialIterationStartValues,

web/src/pages/agent/hooks/use-chat-logic.ts

@@ -0,0 +1,60 @@
+import { MessageType } from '@/constants/chat';
+import { Message } from '@/interfaces/database/chat';
+import { IMessage } from '@/pages/chat/interface';
+import { get } from 'lodash';
+import { useCallback, useMemo } from 'react';
+import { BeginQuery } from '../interface';
+import { buildBeginQueryWithObject } from '../utils';
+type IAwaitCompentData = {
+  derivedMessages: IMessage[];
+  sendFormMessage: (params: {
+    inputs: Record<string, BeginQuery>;
+    id: string;
+  }) => void;
+  canvasId: string;
+};
+const useAwaitCompentData = (props: IAwaitCompentData) => {
+  const { derivedMessages, sendFormMessage, canvasId } = props;
+
+  const getInputs = useCallback((message: Message) => {
+    return get(message, 'data.inputs', {}) as Record<string, BeginQuery>;
+  }, []);
+
+  const buildInputList = useCallback(
+    (message: Message) => {
+      return Object.entries(getInputs(message)).map(([key, val]) => {
+        return {
+          ...val,
+          key,
+        };
+      });
+    },
+    [getInputs],
+  );
+
+  const handleOk = useCallback(
+    (message: Message) => (values: BeginQuery[]) => {
+      const inputs = getInputs(message);
+      const nextInputs = buildBeginQueryWithObject(inputs, values);
+      sendFormMessage({
+        inputs: nextInputs,
+        id: canvasId,
+      });
+    },
+    [getInputs, sendFormMessage, canvasId],
+  );
+
+  const isWaitting = useMemo(() => {
+    const temp = derivedMessages?.some((message, i) => {
+      const flag =
+        message.role === MessageType.Assistant &&
+        derivedMessages.length - 1 === i &&
+        message.data;
+      return flag;
+    });
+    return temp;
+  }, [derivedMessages]);
+  return { getInputs, buildInputList, handleOk, isWaitting };
+};
+
+export { useAwaitCompentData };

web/src/pages/agent/index.tsx

@@ -44,6 +44,7 @@ import {
 } from './hooks/use-save-graph';
 import { useShowEmbedModal } from './hooks/use-show-dialog';
 import { UploadAgentDialog } from './upload-agent-dialog';
+import { useAgentHistoryManager } from './use-agent-history-manager';
 import { VersionDialog } from './version-dialog';
 
 function AgentDropdownMenuItem({
@@ -66,8 +67,7 @@ export default function Agent() {
     showModal: showChatDrawer,
   } = useSetModalState();
   const { t } = useTranslation();
-
-  // const openDocument = useOpenDocument();
+  useAgentHistoryManager();
   const {
     handleExportJson,
     handleImportJson,

web/src/pages/agent/log-sheet/toolTimelineItem.tsx

@@ -14,24 +14,37 @@ import {
 import { cn } from '@/lib/utils';
 import { isEmpty } from 'lodash';
 import { Operator } from '../constant';
-import OperatorIcon from '../operator-icon';
+import OperatorIcon, { SVGIconMap } from '../operator-icon';
 import {
   JsonViewer,
   toLowerCaseStringAndDeleteChar,
   typeMap,
 } from './workFlowTimeline';
-const capitalizeWords = (str: string, separator: string = '_'): string => {
-  if (!str) return '';
+type IToolIcon =
+  | Operator.ArXiv
+  | Operator.GitHub
+  | Operator.Bing
+  | Operator.DuckDuckGo
+  | Operator.Google
+  | Operator.GoogleScholar
+  | Operator.PubMed
+  | Operator.TavilyExtract
+  | Operator.TavilySearch
+  | Operator.Wikipedia
+  | Operator.YahooFinance
+  | Operator.WenCai
+  | Operator.Crawler;
 
-  return str
-    .split(separator)
-    .map((word) => {
-      return word.charAt(0).toUpperCase() + word.slice(1).toLowerCase();
-    })
-    .join(' ');
+const capitalizeWords = (str: string, separator: string = '_'): string[] => {
+  if (!str) return [''];
+
+  const resultStrArr = str.split(separator).map((word) => {
+    return word.charAt(0).toUpperCase() + word.slice(1).toLowerCase();
+  });
+  return resultStrArr;
 };
 const changeToolName = (toolName: any) => {
-  const name = 'Agent ' + capitalizeWords(toolName);
+  const name = 'Agent ' + capitalizeWords(toolName).join(' ');
   return name;
 };
 const ToolTimelineItem = ({
@@ -61,6 +74,8 @@ const ToolTimelineItem = ({
   return (
     <>
       {filteredTools?.map((tool, idx) => {
+        const toolName = capitalizeWords(tool.tool_name, '_').join('');
+
         return (
           <TimelineItem
             key={'tool_' + idx}
@@ -105,7 +120,11 @@ const ToolTimelineItem = ({
                   <div className="size-6 flex items-center justify-center">
                     <OperatorIcon
                       className="size-4"
-                      name={'Agent' as Operator}
+                      name={
+                        (SVGIconMap[toolName as IToolIcon]
+                          ? toolName
+                          : 'Agent') as Operator
+                      }
                     ></OperatorIcon>
                   </div>
                 </div>
@@ -119,12 +138,14 @@ const ToolTimelineItem = ({
                   className="bg-background-card px-3"
                 >
                   <AccordionItem value={idx.toString()}>
-                    <AccordionTrigger>
+                    <AccordionTrigger
+                      hideDownIcon={isShare && isEmpty(tool.arguments)}
+                    >
                       <div className="flex gap-2 items-center">
                         {!isShare && (
                           <span>
                             {parentName(tool.path) + ' '}
-                            {capitalizeWords(tool.tool_name, '_')}
+                            {capitalizeWords(tool.tool_name, '_').join(' ')}
                           </span>
                         )}
                         {isShare && (
@@ -142,7 +163,7 @@ const ToolTimelineItem = ({
                         </span>
                         <span
                           className={cn(
-                            'border-background  -end-1 -top-1 size-2 rounded-full border-2 bg-dot-green',
+                            'border-background  -end-1 -top-1 size-2 rounded-full bg-dot-green',
                           )}
                         >
                           <span className="sr-only">Online</span>
@@ -161,7 +182,7 @@ const ToolTimelineItem = ({
                     )}
                     {isShare && !isEmpty(tool.arguments) && (
                       <AccordionContent>
-                        <div className="space-y-2">
+                        <div className="space-y-2 bg-muted p-2">
                           {tool &&
                             tool.arguments &&
                             Object.entries(tool.arguments).length &&
@@ -171,8 +192,8 @@ const ToolTimelineItem = ({
                                   <div className="text-sm font-medium leading-none">
                                     {key}
                                   </div>
-                                  <div className="text-sm text-muted-foreground">
-                                    {val || ''}
+                                  <div className="text-sm text-muted-foreground mt-1">
+                                    {val as string}
                                   </div>
                                 </div>
                               );

web/src/pages/agent/log-sheet/workFlowTimeline.tsx

@@ -51,7 +51,7 @@ export function JsonViewer({
         src={data}
         displaySize
         collapseStringsAfterLength={100000000000}
-        className="w-full h-[200px] break-words overflow-auto scrollbar-auto p-2 bg-slate-800"
+        className="w-full h-[200px] break-words overflow-auto scrollbar-auto p-2 bg-muted"
       />
     </section>
   );
@@ -81,11 +81,21 @@ export const typeMap = {
   httpRequest: t('flow.logTimeline.httpRequest'),
   wenCai: t('flow.logTimeline.wenCai'),
   yahooFinance: t('flow.logTimeline.yahooFinance'),
+  userFillUp: t('flow.logTimeline.userFillUp'),
 };
 export const toLowerCaseStringAndDeleteChar = (
   str: string,
   char: string = '_',
 ) => str.toLowerCase().replace(/ /g, '').replaceAll(char, '');
+
+// Convert all keys in typeMap to lowercase and output the new typeMap
+export const typeMapLowerCase = Object.fromEntries(
+  Object.entries(typeMap).map(([key, value]) => [
+    toLowerCaseStringAndDeleteChar(key),
+    value,
+  ]),
+);
+
 function getInputsOrOutputs(
   nodeEventList: INodeData[],
   field: 'inputs' | 'outputs',
@@ -247,16 +257,19 @@ export const WorkFlowTimeline = ({
                     className="bg-background-card px-3"
                   >
                     <AccordionItem value={idx.toString()}>
-                      <AccordionTrigger>
+                      <AccordionTrigger
+                        hideDownIcon={isShare && !x.data?.thoughts}
+                      >
                         <div className="flex gap-2 items-center">
                           <span>
                             {!isShare && getNodeName(x.data?.component_name)}
                             {isShare &&
-                              typeMap[
+                              (typeMapLowerCase[
                                 toLowerCaseStringAndDeleteChar(
                                   nodeLabel,
                                 ) as keyof typeof typeMap
-                              ]}
+                              ] ??
+                                nodeLabel)}
                           </span>
                           <span className="text-text-sub-title text-xs">
                             {x.data.elapsed_time?.toString().slice(0, 6)}
@@ -294,7 +307,7 @@ export const WorkFlowTimeline = ({
                       {isShare && x.data?.thoughts && (
                         <AccordionContent>
                           <div className="space-y-2">
-                            <div className="w-full h-[200px] break-words overflow-auto scrollbar-auto p-2 bg-slate-800">
+                            <div className="w-full h-[200px] break-words overflow-auto scrollbar-auto p-2 bg-muted">
                               <HightLightMarkdown>
                                 {x.data.thoughts || ''}
                               </HightLightMarkdown>

web/src/pages/agent/operator-icon.tsx

@@ -38,7 +38,7 @@ export const OperatorIconMap = {
   [Operator.Email]: 'sendemail-0',
 };
 
-const SVGIconMap = {
+export const SVGIconMap = {
   [Operator.ArXiv]: ArxivIcon,
   [Operator.GitHub]: GithubIcon,
   [Operator.Bing]: BingIcon,

web/src/pages/agent/options.ts

@@ -57,6 +57,10 @@ export const LanguageOptions = [
     value: 'de',
     label: 'Deutsch',
   },
+  {
+    value: 'fr',
+    label: 'Français',
+  },
   {
     value: 'et',
     label: 'Eesti',

web/src/pages/agent/use-agent-history-manager.ts

@@ -0,0 +1,163 @@
+import { useEffect, useRef } from 'react';
+import useGraphStore from './store';
+
+// History management class
+export class HistoryManager {
+  private history: { nodes: any[]; edges: any[] }[] = [];
+  private currentIndex: number = -1;
+  private readonly maxSize: number = 50; // Limit maximum number of history records
+  private setNodes: (nodes: any[]) => void;
+  private setEdges: (edges: any[]) => void;
+  private lastSavedState: string = ''; // Used to compare if state has changed
+
+  constructor(
+    setNodes: (nodes: any[]) => void,
+    setEdges: (edges: any[]) => void,
+  ) {
+    this.setNodes = setNodes;
+    this.setEdges = setEdges;
+  }
+
+  // Compare if two states are equal
+  private statesEqual(
+    state1: { nodes: any[]; edges: any[] },
+    state2: { nodes: any[]; edges: any[] },
+  ): boolean {
+    return JSON.stringify(state1) === JSON.stringify(state2);
+  }
+
+  push(nodes: any[], edges: any[]) {
+    const currentState = {
+      nodes: JSON.parse(JSON.stringify(nodes)),
+      edges: JSON.parse(JSON.stringify(edges)),
+    };
+
+    // If state hasn't changed, don't save
+    if (
+      this.history.length > 0 &&
+      this.statesEqual(currentState, this.history[this.currentIndex])
+    ) {
+      return;
+    }
+
+    // If current index is not at the end of history, remove subsequent states
+    if (this.currentIndex < this.history.length - 1) {
+      this.history.splice(this.currentIndex + 1);
+    }
+
+    // Add current state
+    this.history.push(currentState);
+
+    // Limit history record size
+    if (this.history.length > this.maxSize) {
+      this.history.shift();
+      this.currentIndex = this.history.length - 1;
+    } else {
+      this.currentIndex = this.history.length - 1;
+    }
+
+    // Update last saved state
+    this.lastSavedState = JSON.stringify(currentState);
+  }
+
+  undo() {
+    if (this.canUndo()) {
+      this.currentIndex--;
+      const prevState = this.history[this.currentIndex];
+      this.setNodes(JSON.parse(JSON.stringify(prevState.nodes)));
+      this.setEdges(JSON.parse(JSON.stringify(prevState.edges)));
+      return true;
+    }
+    return false;
+  }
+
+  redo() {
+    console.log('redo');
+    if (this.canRedo()) {
+      this.currentIndex++;
+      const nextState = this.history[this.currentIndex];
+      this.setNodes(JSON.parse(JSON.stringify(nextState.nodes)));
+      this.setEdges(JSON.parse(JSON.stringify(nextState.edges)));
+      return true;
+    }
+    return false;
+  }
+
+  canUndo() {
+    return this.currentIndex > 0;
+  }
+
+  canRedo() {
+    return this.currentIndex < this.history.length - 1;
+  }
+
+  // Reset history records
+  reset() {
+    this.history = [];
+    this.currentIndex = -1;
+    this.lastSavedState = '';
+  }
+}
+
+export const useAgentHistoryManager = () => {
+  // Get current state and history state
+  const nodes = useGraphStore((state) => state.nodes);
+  const edges = useGraphStore((state) => state.edges);
+  const setNodes = useGraphStore((state) => state.setNodes);
+  const setEdges = useGraphStore((state) => state.setEdges);
+
+  // Use useRef to keep HistoryManager instance unchanged
+  const historyManagerRef = useRef<HistoryManager | null>(null);
+
+  // Initialize HistoryManager
+  if (!historyManagerRef.current) {
+    historyManagerRef.current = new HistoryManager(setNodes, setEdges);
+  }
+
+  const historyManager = historyManagerRef.current;
+
+  // Save state history - use useEffect instead of useMemo to avoid re-rendering
+  useEffect(() => {
+    historyManager.push(nodes, edges);
+  }, [nodes, edges, historyManager]);
+
+  // Keyboard event handling
+  useEffect(() => {
+    const handleKeyDown = (e: KeyboardEvent) => {
+      // Check if focused on an input element
+      const activeElement = document.activeElement;
+      const isInputFocused =
+        activeElement instanceof HTMLInputElement ||
+        activeElement instanceof HTMLTextAreaElement ||
+        activeElement?.hasAttribute('contenteditable');
+
+      // Skip keyboard shortcuts if typing in an input field
+      if (isInputFocused) {
+        return;
+      }
+      // Ctrl+Z or Cmd+Z undo
+      if (
+        (e.ctrlKey || e.metaKey) &&
+        (e.key === 'z' || e.key === 'Z') &&
+        !e.shiftKey
+      ) {
+        e.preventDefault();
+        historyManager.undo();
+      }
+      // Ctrl+Shift+Z or Cmd+Shift+Z redo
+      else if (
+        (e.ctrlKey || e.metaKey) &&
+        (e.key === 'z' || e.key === 'Z') &&
+        e.shiftKey
+      ) {
+        e.preventDefault();
+        historyManager.redo();
+      }
+    };
+
+    document.addEventListener('keydown', handleKeyDown);
+    return () => {
+      document.removeEventListener('keydown', handleKeyDown);
+    };
+  }, [historyManager]);
+};

web/src/pages/agents/agent-log-page.tsx

@@ -231,7 +231,7 @@ const AgentLogPage: React.FC = () => {
         <div className="flex justify-between items-center">
           <h1 className="text-2xl font-bold mb-4">Log</h1>
 
-          <div className="flex justify-end space-x-2 mb-4">
+          <div className="flex justify-end space-x-2 mb-4 text-foreground">
             <div className="flex items-center space-x-2">
               <span>ID/Title</span>
               <SearchInput

web/src/pages/agents/index.tsx

@@ -10,7 +10,7 @@ import { useCallback } from 'react';
 import { AgentCard } from './agent-card';
 import { useRenameAgent } from './use-rename-agent';
 
-export default function Agent() {
+export default function Agents() {
   const { data, pagination, setPagination, searchString, handleInputChange } =
     useFetchAgentListByPage();
   const { navigateToAgentTemplates } = useNavigatePage();

web/src/pages/chat/chat-container/index.tsx

@@ -1,7 +1,6 @@
 import MessageItem from '@/components/message-item';
 import { MessageType } from '@/constants/chat';
 import { Flex, Spin } from 'antd';
-import { useRef } from 'react';
 import {
   useCreateConversationBeforeUploadDocument,
   useGetFileIcon,
@@ -19,7 +18,6 @@ import {
   useFetchNextDialog,
   useGetChatSearchParams,
 } from '@/hooks/chat-hooks';
-import { useScrollToBottom } from '@/hooks/logic-hooks';
 import { useFetchUserInfo } from '@/hooks/user-setting-hooks';
 import { buildMessageUuidWithRole } from '@/utils/chat';
 import { memo } from 'react';
@@ -34,10 +32,10 @@ const ChatContainer = ({ controller }: IProps) => {
   const { data: conversation } = useFetchNextConversation();
   const { data: currentDialog } = useFetchNextDialog();
 
-  const messageContainerRef = useRef<HTMLDivElement>(null);
   const {
     value,
-    ref,
+    scrollRef,
+    messageContainerRef,
     loading,
     sendLoading,
     derivedMessages,
@@ -47,10 +45,6 @@ const ChatContainer = ({ controller }: IProps) => {
     removeMessageById,
     stopOutputMessage,
   } = useSendNextMessage(controller);
-  const { scrollRef, isAtBottom, scrollToBottom } = useScrollToBottom(
-    derivedMessages,
-    messageContainerRef,
-  );
 
   const { visible, hideModal, documentId, selectedChunk, clickDocumentButton } =
     useClickDrawer();
@@ -61,11 +55,6 @@ const ChatContainer = ({ controller }: IProps) => {
   const { createConversationBeforeUploadDocument } =
     useCreateConversationBeforeUploadDocument();
 
-  const handleSend = (msg) => {
-    // your send logic
-    setTimeout(scrollToBottom, 0);
-  };
-
   return (
     <>
       <Flex flex={1} className={styles.chatContainer} vertical>

web/src/pages/chat/hooks.ts

@@ -291,7 +291,8 @@ export const useSetConversation = () => {
 
 export const useSelectNextMessages = () => {
   const {
-    ref,
+    scrollRef,
+    messageContainerRef,
     setDerivedMessages,
     derivedMessages,
     addNewestAnswer,
@@ -335,7 +336,8 @@ export const useSelectNextMessages = () => {
   }, [conversation.message, conversationId, setDerivedMessages, isNew]);
 
   return {
-    ref,
+    scrollRef,
+    messageContainerRef,
     derivedMessages,
     loading,
     addNewestAnswer,
@@ -371,7 +373,8 @@ export const useSendNextMessage = (controller: AbortController) => {
     api.completeConversation,
   );
   const {
-    ref,
+    scrollRef,
+    messageContainerRef,
     derivedMessages,
     loading,
     addNewestAnswer,
@@ -499,7 +502,8 @@ export const useSendNextMessage = (controller: AbortController) => {
     regenerateMessage,
     sendLoading: !done,
     loading,
-    ref,
+    scrollRef,
+    messageContainerRef,
     derivedMessages,
     removeMessageById,
     stopOutputMessage,

web/src/pages/datasets/index.tsx

@@ -53,7 +53,7 @@ export default function Datasets() {
   );
 
   return (
-    <section className="py-4 text-foreground">
+    <section className="py-4 flex-1 flex flex-col">
       <ListFilterBar
         title={t('header.knowledgeBase')}
         searchString={searchString}
@@ -69,16 +69,18 @@ export default function Datasets() {
           {t('knowledgeList.createKnowledgeBase')}
         </Button>
       </ListFilterBar>
-      <div className="flex flex-wrap gap-4 max-h-[78vh] overflow-auto px-8">
-        {kbs.map((dataset) => {
-          return (
-            <DatasetCard
-              dataset={dataset}
-              key={dataset.id}
-              showDatasetRenameModal={showDatasetRenameModal}
-            ></DatasetCard>
-          );
-        })}
+      <div className="flex-1">
+        <div className="flex flex-wrap gap-4 max-h-[78vh] overflow-auto px-8">
+          {kbs.map((dataset) => {
+            return (
+              <DatasetCard
+                dataset={dataset}
+                key={dataset.id}
+                showDatasetRenameModal={showDatasetRenameModal}
+              ></DatasetCard>
+            );
+          })}
+        </div>
       </div>
       <div className="mt-8 px-8">
         <RAGFlowPagination

web/src/pages/flow/constant.tsx

@@ -835,6 +835,10 @@ export const LanguageOptions = [
     value: 'de',
     label: 'Deutsch',
   },
+  {
+    value: 'fr',
+    label: 'Français',
+  },
   {
     value: 'et',
     label: 'Eesti',

web/src/pages/home/chat-list.tsx

@@ -2,9 +2,9 @@ import { useFetchDialogList } from '@/hooks/use-chat-request';
 import { ApplicationCard } from './application-card';
 
 export function ChatList() {
-  const { data } = useFetchDialogList(true);
+  const { data } = useFetchDialogList();
 
-  return data
+  return data.dialogs
     .slice(0, 10)
     .map((x) => (
       <ApplicationCard

web/src/pages/next-chats/chat-card.tsx

@@ -1,52 +1,45 @@
-import { Avatar, AvatarFallback, AvatarImage } from '@/components/ui/avatar';
-import { Button } from '@/components/ui/button';
+import { MoreButton } from '@/components/more-button';
+import { RAGFlowAvatar } from '@/components/ragflow-avatar';
 import { Card, CardContent } from '@/components/ui/card';
 import { useNavigatePage } from '@/hooks/logic-hooks/navigate-hooks';
 import { IDialog } from '@/interfaces/database/chat';
-import { formatPureDate } from '@/utils/date';
-import { ChevronRight, Trash2 } from 'lucide-react';
+import { formatDate } from '@/utils/date';
+import { ChatDropdown } from './chat-dropdown';
+import { useRenameChat } from './hooks/use-rename-chat';
 
-interface IProps {
+export type IProps = {
   data: IDialog;
-}
+} & Pick<ReturnType<typeof useRenameChat>, 'showChatRenameModal'>;
 
-export function ChatCard({ data }: IProps) {
+export function ChatCard({ data, showChatRenameModal }: IProps) {
   const { navigateToChat } = useNavigatePage();
 
   return (
-    <Card className="bg-colors-background-inverse-weak  border-colors-outline-neutral-standard">
-      <CardContent className="p-4">
-        <div className="flex justify-between mb-4">
-          {data.icon ? (
-            <div
-              className="w-[70px] h-[70px] rounded-xl bg-cover"
-              style={{ backgroundImage: `url(${data.icon})` }}
-            />
-          ) : (
-            <Avatar className="w-[70px] h-[70px]">
-              <AvatarImage src="https://github.com/shadcn.png" />
-              <AvatarFallback>CN</AvatarFallback>
-            </Avatar>
-          )}
-        </div>
-        <h3 className="text-xl font-bold mb-2">{data.name}</h3>
-        <p>An app that does things An app that does things</p>
-        <section className="flex justify-between pt-3">
-          <div>
-            Search app
-            <p className="text-sm opacity-80">
-              {formatPureDate(data.update_time)}
+    <Card key={data.id} className="w-40" onClick={navigateToChat(data.id)}>
+      <CardContent className="p-2.5 pt-2 group">
+        <section className="flex justify-between mb-2">
+          <div className="flex gap-2 items-center">
+            <RAGFlowAvatar
+              className="size-6 rounded-lg"
+              avatar={data.icon}
+              name={data.name || 'CN'}
+            ></RAGFlowAvatar>
+          </div>
+          <ChatDropdown chat={data} showChatRenameModal={showChatRenameModal}>
+            <MoreButton></MoreButton>
+          </ChatDropdown>
+        </section>
+        <div className="flex justify-between items-end">
+          <div className="w-full">
+            <h3 className="text-lg font-semibold mb-2 line-clamp-1">
+              {data.name}
+            </h3>
+            <p className="text-xs text-text-sub-title">{data.description}</p>
+            <p className="text-xs text-text-sub-title">
+              {formatDate(data.update_time)}
             </p>
           </div>
-          <div className="space-x-2">
-            <Button variant="icon" size="icon" onClick={navigateToChat}>
-              <ChevronRight className="h-6 w-6" />
-            </Button>
-            <Button variant="icon" size="icon">
-              <Trash2 />
-            </Button>
-          </div>
-        </section>
+        </div>
       </CardContent>
     </Card>
   );

web/src/pages/next-chats/chat-dropdown.tsx

@@ -0,0 +1,64 @@
+import { ConfirmDeleteDialog } from '@/components/confirm-delete-dialog';
+import {
+  DropdownMenu,
+  DropdownMenuContent,
+  DropdownMenuItem,
+  DropdownMenuSeparator,
+  DropdownMenuTrigger,
+} from '@/components/ui/dropdown-menu';
+import { useRemoveDialog } from '@/hooks/use-chat-request';
+import { IDialog } from '@/interfaces/database/chat';
+import { PenLine, Trash2 } from 'lucide-react';
+import { MouseEventHandler, PropsWithChildren, useCallback } from 'react';
+import { useTranslation } from 'react-i18next';
+import { useRenameChat } from './hooks/use-rename-chat';
+
+export function ChatDropdown({
+  children,
+  showChatRenameModal,
+  chat,
+}: PropsWithChildren &
+  Pick<ReturnType<typeof useRenameChat>, 'showChatRenameModal'> & {
+    chat: IDialog;
+  }) {
+  const { t } = useTranslation();
+  const { removeDialog } = useRemoveDialog();
+
+  const handleShowChatRenameModal: MouseEventHandler<HTMLDivElement> =
+    useCallback(
+      (e) => {
+        e.stopPropagation();
+        showChatRenameModal(chat);
+      },
+      [chat, showChatRenameModal],
+    );
+
+  const handleDelete: MouseEventHandler<HTMLDivElement> = useCallback(() => {
+    removeDialog([chat.id]);
+  }, [chat.id, removeDialog]);
+
+  return (
+    <DropdownMenu>
+      <DropdownMenuTrigger asChild>{children}</DropdownMenuTrigger>
+      <DropdownMenuContent>
+        <DropdownMenuItem onClick={handleShowChatRenameModal}>
+          {t('common.rename')} <PenLine />
+        </DropdownMenuItem>
+        <DropdownMenuSeparator />
+        <ConfirmDeleteDialog onOk={handleDelete}>
+          <DropdownMenuItem
+            className="text-text-delete-red"
+            onSelect={(e) => {
+              e.preventDefault();
+            }}
+            onClick={(e) => {
+              e.stopPropagation();
+            }}
+          >
+            {t('common.delete')} <Trash2 />
+          </DropdownMenuItem>
+        </ConfirmDeleteDialog>
+      </DropdownMenuContent>
+    </DropdownMenu>
+  );
+}

web/src/pages/next-chats/chat/index.tsx

@@ -1,6 +1,7 @@
 import { PageHeader } from '@/components/page-header';
 import { Button } from '@/components/ui/button';
 import { useNavigatePage } from '@/hooks/logic-hooks/navigate-hooks';
+import { useFetchDialog } from '@/hooks/use-chat-request';
 import { EllipsisVertical } from 'lucide-react';
 import { AppSettings } from './app-settings';
 import { ChatBox } from './chat-box';
@@ -8,10 +9,11 @@ import { Sessions } from './sessions';
 
 export default function Chat() {
   const { navigateToChatList } = useNavigatePage();
+  useFetchDialog();
 
   return (
     <section className="h-full flex flex-col">
-      <PageHeader back={navigateToChatList} title="Chat app 01">
+      <PageHeader>
         <div className="flex items-center gap-2">
           <Button variant={'icon'} size={'icon'}>
             <EllipsisVertical />

web/src/pages/next-chats/hooks/use-rename-chat.ts

@@ -0,0 +1,78 @@
+import { useSetModalState } from '@/hooks/common-hooks';
+import { useSetDialog } from '@/hooks/use-chat-request';
+import { IDialog } from '@/interfaces/database/chat';
+import { isEmpty } from 'lodash';
+import { useCallback, useState } from 'react';
+
+const InitialData = {
+  name: '',
+  icon: '',
+  language: 'English',
+  prompt_config: {
+    empty_response: '',
+    prologue: '你好！ 我是你的助理，有什么可以帮到你的吗？',
+    quote: true,
+    keyword: false,
+    tts: false,
+    system:
+      '你是一个智能助手，请总结知识库的内容来回答问题，请列举知识库中的数据详细回答。当所有知识库内容都与问题无关时，你的回答必须包括“知识库中未找到您要的答案！”这句话。回答需要考虑聊天历史。\n        以下是知识库：\n        {knowledge}\n        以上是知识库。',
+    refine_multiturn: false,
+    use_kg: false,
+    reasoning: false,
+    parameters: [{ key: 'knowledge', optional: false }],
+  },
+  llm_id: '',
+  llm_setting: {},
+  similarity_threshold: 0.2,
+  vector_similarity_weight: 0.30000000000000004,
+  top_n: 8,
+};
+
+export const useRenameChat = () => {
+  const [chat, setChat] = useState<IDialog>({} as IDialog);
+  const {
+    visible: chatRenameVisible,
+    hideModal: hideChatRenameModal,
+    showModal: showChatRenameModal,
+  } = useSetModalState();
+  const { setDialog, loading } = useSetDialog();
+
+  const onChatRenameOk = useCallback(
+    async (name: string) => {
+      const nextChat = {
+        ...(isEmpty(chat) ? InitialData : chat),
+        name,
+      };
+      const ret = await setDialog(nextChat);
+
+      if (ret === 0) {
+        hideChatRenameModal();
+      }
+    },
+    [setDialog, chat, hideChatRenameModal],
+  );
+
+  const handleShowChatRenameModal = useCallback(
+    (record?: IDialog) => {
+      if (record) {
+        setChat(record);
+      }
+      showChatRenameModal();
+    },
+    [showChatRenameModal],
+  );
+
+  const handleHideModal = useCallback(() => {
+    hideChatRenameModal();
+    setChat({} as IDialog);
+  }, [hideChatRenameModal]);
+
+  return {
+    chatRenameLoading: loading,
+    initialChatName: chat?.name,
+    onChatRenameOk,
+    chatRenameVisible,
+    hideChatRenameModal: handleHideModal,
+    showChatRenameModal: handleShowChatRenameModal,
+  };
+};

web/src/pages/next-chats/index.tsx

@@ -1,25 +1,77 @@
 import ListFilterBar from '@/components/list-filter-bar';
+import { RenameDialog } from '@/components/rename-dialog';
 import { Button } from '@/components/ui/button';
-import { useFetchChatAppList } from '@/hooks/chat-hooks';
+import { RAGFlowPagination } from '@/components/ui/ragflow-pagination';
+import { useFetchDialogList } from '@/hooks/use-chat-request';
+import { pick } from 'lodash';
 import { Plus } from 'lucide-react';
+import { useCallback } from 'react';
+import { useTranslation } from 'react-i18next';
 import { ChatCard } from './chat-card';
+import { useRenameChat } from './hooks/use-rename-chat';
 
 export default function ChatList() {
-  const { data: chatList } = useFetchChatAppList();
+  const { data, setPagination, pagination } = useFetchDialogList();
+  const { t } = useTranslation();
+  const {
+    initialChatName,
+    chatRenameVisible,
+    showChatRenameModal,
+    hideChatRenameModal,
+    onChatRenameOk,
+    chatRenameLoading,
+  } = useRenameChat();
+
+  const handlePageChange = useCallback(
+    (page: number, pageSize?: number) => {
+      setPagination({ page, pageSize });
+    },
+    [setPagination],
+  );
+
+  const handleShowCreateModal = useCallback(() => {
+    showChatRenameModal();
+  }, [showChatRenameModal]);
 
   return (
-    <section className="p-8">
-      <ListFilterBar title="Chat apps">
-        <Button variant={'tertiary'} size={'sm'}>
-          <Plus className="mr-2 h-4 w-4" />
-          Create app
-        </Button>
-      </ListFilterBar>
-      <div className="grid gap-6 sm:grid-cols-1 md:grid-cols-2 lg:grid-cols-4 xl:grid-cols-6 2xl:grid-cols-8">
-        {chatList.map((x) => {
-          return <ChatCard key={x.id} data={x}></ChatCard>;
-        })}
+    <section className="flex flex-col w-full flex-1">
+      <div className="px-8 pt-8">
+        <ListFilterBar title="Chat apps">
+          <Button onClick={handleShowCreateModal}>
+            <Plus className="size-2.5" />
+            {t('chat.createChat')}
+          </Button>
+        </ListFilterBar>
       </div>
+      <div className="flex-1 overflow-auto">
+        <div className="flex flex-wrap gap-4 px-8">
+          {data.dialogs.map((x) => {
+            return (
+              <ChatCard
+                key={x.id}
+                data={x}
+                showChatRenameModal={showChatRenameModal}
+              ></ChatCard>
+            );
+          })}
+        </div>
+      </div>
+      <div className="mt-8 px-8 pb-8">
+        <RAGFlowPagination
+          {...pick(pagination, 'current', 'pageSize')}
+          total={pagination.total}
+          onChange={handlePageChange}
+        ></RAGFlowPagination>
+      </div>
+      {chatRenameVisible && (
+        <RenameDialog
+          hideModal={hideChatRenameModal}
+          onOk={onChatRenameOk}
+          initialName={initialChatName}
+          loading={chatRenameLoading}
+          title={initialChatName || t('chat.createChat')}
+        ></RenameDialog>
+      )}
     </section>
   );
 }