Refactor:Improve VoyageRerank not texts handling (#9539)

### What problem does this PR solve?

Improve VoyageRerank not texts handling

### Type of change

- [x] Refactoring

.github/ISSUE_TEMPLATE/agent_scenario_request.yml

@@ -0,0 +1,46 @@
+name: "❤️‍🔥ᴬᴳᴱᴺᵀ Agent scenario request"
+description: Propose a agent scenario request for RAGFlow.
+title: "[Agent Scenario Request]: "
+labels: ["❤️‍🔥ᴬᴳᴱᴺᵀ agent scenario"]
+body:
+  - type: checkboxes
+    attributes:
+      label: Self Checks
+      description: "Please check the following in order to be responded in time :)"
+      options:
+        - label: I have searched for existing issues [search for existing issues](https://github.com/infiniflow/ragflow/issues), including closed ones.
+          required: true
+        - label: I confirm that I am using English to submit this report ([Language Policy](https://github.com/infiniflow/ragflow/issues/5910)).
+          required: true
+        - label: Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) ([Language Policy](https://github.com/infiniflow/ragflow/issues/5910)).
+          required: true
+        - label: "Please do not modify this template :) and fill in all the required fields."
+          required: true
+  - type: textarea
+    attributes:
+      label: Is your feature request related to a scenario?
+      description: |
+        A clear and concise description of what the scenario is. Ex. I'm always frustrated when [...]
+      render: Markdown
+    validations:
+      required: false
+  - type: textarea
+    attributes:
+      label: Describe the feature you'd like
+      description: A clear and concise description of what you want to happen.
+    validations:
+      required: true
+  - type: textarea
+    attributes:
+      label: Documentation, adoption, use case
+      description: If you can, explain some scenarios how users might use this, situations it would be helpful in. Any API designs, mockups, or diagrams are also helpful.
+      render: Markdown
+    validations:
+      required: false
+  - type: textarea
+    attributes:
+      label: Additional information
+      description: |
+        Add any other context or screenshots about the feature request here.
+    validations:
+      required: false

agent/canvas.py

@@ -484,7 +484,7 @@ class Canvas:
             threads.append(exe.submit(FileService.parse, file["name"], FileService.get_blob(file["created_by"], file["id"]), True, file["created_by"]))
         return [th.result() for th in threads]
 
-    def tool_use_callback(self, agent_id: str, func_name: str, params: dict, result: Any):
+    def tool_use_callback(self, agent_id: str, func_name: str, params: dict, result: Any, elapsed_time=None):
         agent_ids = agent_id.split("-->")
         agent_name = self.get_component_name(agent_ids[0])
         path = agent_name if len(agent_ids) < 2 else agent_name+"-->"+"-->".join(agent_ids[1:])
@@ -493,16 +493,16 @@ class Canvas:
             if bin:
                 obj = json.loads(bin.encode("utf-8"))
                 if obj[-1]["component_id"] == agent_ids[0]:
-                    obj[-1]["trace"].append({"path": path, "tool_name": func_name, "arguments": params, "result": result})
+                    obj[-1]["trace"].append({"path": path, "tool_name": func_name, "arguments": params, "result": result, "elapsed_time": elapsed_time})
                 else:
                     obj.append({
                     "component_id": agent_ids[0],
-                    "trace": [{"path": path, "tool_name": func_name, "arguments": params, "result": result}]
+                    "trace": [{"path": path, "tool_name": func_name, "arguments": params, "result": result, "elapsed_time": elapsed_time}]
                 })
             else:
                 obj = [{
                     "component_id": agent_ids[0],
-                    "trace": [{"path": path, "tool_name": func_name, "arguments": params, "result": result}]
+                    "trace": [{"path": path, "tool_name": func_name, "arguments": params, "result": result, "elapsed_time": elapsed_time}]
                 }]
             REDIS_CONN.set_obj(f"{self.task_id}-{self.message_id}-logs", obj, 60*10)
         except Exception as e:

agent/component/agent_with_tools.py

@@ -22,9 +22,10 @@ from functools import partial
 from typing import Any
 
 import json_repair
-
+from timeit import default_timer as timer
 from agent.tools.base import LLMToolPluginCallSession, ToolParamBase, ToolBase, ToolMeta
-from api.db.services.llm_service import LLMBundle, TenantLLMService
+from api.db.services.llm_service import LLMBundle
+from api.db.services.tenant_llm_service import TenantLLMService
 from api.db.services.mcp_server_service import MCPServerService
 from api.utils.api_utils import timeout
 from rag.prompts import message_fit_in
@@ -165,7 +166,7 @@ class Agent(LLM, ToolBase):
         _, msg = message_fit_in([{"role": "system", "content": prompt}, *msg], int(self.chat_mdl.max_length * 0.97))
         use_tools = []
         ans = ""
-        for delta_ans, tk in self._react_with_tools_streamly(msg, use_tools):
+        for delta_ans, tk in self._react_with_tools_streamly(prompt, msg, use_tools):
             ans += delta_ans
 
         if ans.find("**ERROR**") >= 0:
@@ -185,7 +186,7 @@ class Agent(LLM, ToolBase):
         _, msg = message_fit_in([{"role": "system", "content": prompt}, *msg], int(self.chat_mdl.max_length * 0.97))
         answer_without_toolcall = ""
         use_tools = []
-        for delta_ans,_ in self._react_with_tools_streamly(msg, use_tools):
+        for delta_ans,_ in self._react_with_tools_streamly(prompt, msg, use_tools):
             if delta_ans.find("**ERROR**") >= 0:
                 if self.get_exception_default_value():
                     self.set_output("content", self.get_exception_default_value())
@@ -208,20 +209,21 @@ class Agent(LLM, ToolBase):
                                                   ]):
             yield delta_ans
 
-    def _react_with_tools_streamly(self, history: list[dict], use_tools):
+    def _react_with_tools_streamly(self, prompt, history: list[dict], use_tools):
         token_count = 0
         tool_metas = self.tool_meta
         hist = deepcopy(history)
         last_calling = ""
         if len(hist) > 3:
+            st = timer()
             user_request = full_question(messages=history, chat_mdl=self.chat_mdl)
-            self.callback("Multi-turn conversation optimization", {}, user_request)
+            self.callback("Multi-turn conversation optimization", {}, user_request, elapsed_time=timer()-st)
         else:
             user_request = history[-1]["content"]
 
         def use_tool(name, args):
             nonlocal hist, use_tools, token_count,last_calling,user_request
-            print(f"{last_calling=} == {name=}", )
+            logging.info(f"{last_calling=} == {name=}")
             # Summarize of function calling
             #if all([
             #    isinstance(self.toolcall_session.get_tool_obj(name), Agent),
@@ -243,7 +245,7 @@ class Agent(LLM, ToolBase):
 
         def complete():
             nonlocal hist
-            need2cite = self._canvas.get_reference()["chunks"] and self._id.find("-->") < 0
+            need2cite = self._param.cite and self._canvas.get_reference()["chunks"] and self._id.find("-->") < 0
             cited = False
             if hist[0]["role"] == "system" and need2cite:
                 if len(hist) < 7:
@@ -262,12 +264,13 @@ class Agent(LLM, ToolBase):
             if not need2cite or cited:
                 return
 
+            st = timer()
             txt = ""
             for delta_ans in self._gen_citations(entire_txt):
                 yield delta_ans, 0
                 txt += delta_ans
 
-            self.callback("gen_citations", {}, txt)
+            self.callback("gen_citations", {}, txt, elapsed_time=timer()-st)
 
         def append_user_content(hist, content):
             if hist[-1]["role"] == "user":
@@ -275,8 +278,9 @@ class Agent(LLM, ToolBase):
             else:
                 hist.append({"role": "user", "content": content})
 
-        task_desc = analyze_task(self.chat_mdl, user_request, tool_metas)
-        self.callback("analyze_task", {}, task_desc)
+        st = timer()
+        task_desc = analyze_task(self.chat_mdl, prompt, user_request, tool_metas)
+        self.callback("analyze_task", {}, task_desc, elapsed_time=timer()-st)
         for _ in range(self._param.max_rounds + 1):
             response, tk = next_step(self.chat_mdl, hist, tool_metas, task_desc)
             # self.callback("next_step", {}, str(response)[:256]+"...")
@@ -302,9 +306,10 @@ class Agent(LLM, ToolBase):
 
                         thr.append(executor.submit(use_tool, name, args))
 
+                    st = timer()
                     reflection = reflect(self.chat_mdl, hist, [th.result() for th in thr])
                     append_user_content(hist, reflection)
-                    self.callback("reflection", {}, str(reflection))
+                    self.callback("reflection", {}, str(reflection), elapsed_time=timer()-st)
 
             except Exception as e:
                 logging.exception(msg=f"Wrong JSON argument format in LLM ReAct response: {e}")

agent/component/base.py

@@ -529,8 +529,12 @@ class ComponentBase(ABC):
     @staticmethod
     def string_format(content: str, kv: dict[str, str]) -> str:
         for n, v in kv.items():
+            def repl(_match, val=v):
+                return str(val) if val is not None else ""
             content = re.sub(
-                r"\{%s\}" % re.escape(n), v, content
+                r"\{%s\}" % re.escape(n),
+                repl,
+                content
             )
         return content
 

agent/component/begin.py

@@ -39,7 +39,10 @@ class Begin(UserFillUp):
     def _invoke(self, **kwargs):
         for k, v in kwargs.get("inputs", {}).items():
             if isinstance(v, dict) and v.get("type", "").lower().find("file") >=0:
-                v = self._canvas.get_files([v["value"]])
+                if v.get("optional") and v.get("value", None) is None:
+                    v = None
+                else:
+                    v = self._canvas.get_files([v["value"]])
             else:
                 v = v.get("value")
             self.set_output(k, v)

agent/component/invoke.py

@@ -57,7 +57,7 @@ class Invoke(ComponentBase, ABC):
     def _invoke(self, **kwargs):
         args = {}
         for para in self._param.variables:
-            if para.get("value") is not None:
+            if para.get("value"):
                 args[para["key"]] = para["value"]
             else:
                 args[para["key"]] = self._canvas.get_variable_value(para["ref"])
@@ -139,4 +139,4 @@ class Invoke(ComponentBase, ABC):
         assert False, self.output()
 
     def thoughts(self) -> str:
-        return "Waiting for the server respond..."
+        return "Waiting for the server respond..."

agent/component/llm.py

@@ -17,14 +17,15 @@ import json
 import logging
 import os
 import re
-from typing import Any
+from typing import Any, Generator
 
 import json_repair
 from copy import deepcopy
 from functools import partial
 
 from api.db import LLMType
-from api.db.services.llm_service import LLMBundle, TenantLLMService
+from api.db.services.llm_service import LLMBundle
+from api.db.services.tenant_llm_service import TenantLLMService
 from agent.component.base import ComponentBase, ComponentParamBase
 from api.utils.api_utils import timeout
 from rag.prompts import message_fit_in, citation_prompt
@@ -144,7 +145,7 @@ class LLM(ComponentBase):
         prompt = self.string_format(prompt, args)
         for m in msg:
             m["content"] = self.string_format(m["content"], args)
-        if self._canvas.get_reference()["chunks"]:
+        if self._param.cite and self._canvas.get_reference()["chunks"]:
             prompt += citation_prompt()
 
         return prompt, msg
@@ -154,7 +155,7 @@ class LLM(ComponentBase):
             return self.chat_mdl.chat(msg[0]["content"], msg[1:], self._param.gen_conf(), **kwargs)
         return self.chat_mdl.chat(msg[0]["content"], msg[1:], self._param.gen_conf(), images=self.imgs, **kwargs)
 
-    def _generate_streamly(self, msg:list[dict], **kwargs) -> str:
+    def _generate_streamly(self, msg:list[dict], **kwargs) -> Generator[str, None, None]:
         ans = ""
         last_idx = 0
         endswith_think = False

agent/templates/knowledge_base_report.json

@@ -0,0 +1,327 @@
+{
+    "id": 20,
+    "title": "Report Agent Using Knowledge Base",
+    "description": "A report generation assistant using local knowledge base, with advanced capabilities in task planning, reasoning, and reflective analysis. Recommended for academic research paper Q&A",
+    "canvas_type": "Agent",
+    "dsl": {
+        "components": {
+            "Agent:NewPumasLick": {
+                "downstream": [
+                    "Message:OrangeYearsShine"
+                ],
+                "obj": {
+                    "component_name": "Agent",
+                    "params": {
+                        "delay_after_error": 1,
+                        "description": "",
+                        "exception_comment": "",
+                        "exception_default_value": "",
+                        "exception_goto": [],
+                        "exception_method": null,
+                        "frequencyPenaltyEnabled": false,
+                        "frequency_penalty": 0.5,
+                        "llm_id": "qwen3-235b-a22b-instruct-2507@Tongyi-Qianwen",
+                        "maxTokensEnabled": true,
+                        "max_retries": 3,
+                        "max_rounds": 3,
+                        "max_tokens": 128000,
+                        "mcp": [],
+                        "message_history_window_size": 12,
+                        "outputs": {
+                            "content": {
+                                "type": "string",
+                                "value": ""
+                            }
+                        },
+                        "parameter": "Precise",
+                        "presencePenaltyEnabled": false,
+                        "presence_penalty": 0.5,
+                        "prompts": [
+                            {
+                                "content": "# User Query\n {sys.query}",
+                                "role": "user"
+                            }
+                        ],
+                        "sys_prompt": "## Role & Task\nYou are a **\u201cKnowledge Base Retrieval Q\\&A Agent\u201d** whose goal is to break down the user\u2019s question into retrievable subtasks, and then produce a multi-source-verified, structured, and actionable research report using the internal knowledge base.\n## Execution Framework (Detailed Steps & Key Points)\n1. **Assessment & Decomposition**\n   * Actions:\n     * Automatically extract: main topic, subtopics, entities (people/organizations/products/technologies), time window, geographic/business scope.\n     * Output as a list: N facts/data points that must be collected (*N* ranges from 5\u201320 depending on question complexity).\n2. **Query Type Determination (Rule-Based)**\n   * Example rules:\n     * If the question involves a single issue but requests \u201cmethod comparison/multiple explanations\u201d \u2192 use **depth-first**.\n     * If the question can naturally be split into \u22653 independent sub-questions \u2192 use **breadth-first**.\n     * If the question can be answered by a single fact/specification/definition \u2192 use **simple query**.\n3. **Research Plan Formulation**\n   * Depth-first: define 3\u20135 perspectives (methodology/stakeholders/time dimension/technical route, etc.), assign search keywords, target document types, and output format for each perspective.\n   * Breadth-first: list subtasks, prioritize them, and assign search terms.\n   * Simple query: directly provide the search sentence and required fields.\n4. **Retrieval Execution**\n   * After retrieval: perform coverage check (does it contain the key facts?) and quality check (source diversity, authority, latest update time).\n   * If standards are not met, automatically loop: rewrite queries (synonyms/cross-domain terms) and retry \u22643 times, or flag as requiring external search.\n5. **Integration & Reasoning**\n   * Build the answer using a **fact\u2013evidence\u2013reasoning** chain. For each conclusion, attach 1\u20132 strongest pieces of evidence.\n---\n## Quality Gate Checklist (Verify at Each Stage)\n* **Stage 1 (Decomposition)**:\n  * [ ] Key concepts and expected outputs identified\n  * [ ] Required facts/data points listed\n* **Stage 2 (Retrieval)**:\n  * [ ] Meets quality standards (see above)\n  * [ ] If not met: execute query iteration\n* **Stage 3 (Generation)**:\n  * [ ] Each conclusion has at least one direct evidence source\n  * [ ] State assumptions/uncertainties\n  * [ ] Provide next-step suggestions or experiment/retrieval plans\n  * [ ] Final length and depth match user expectations (comply with word count/format if specified)\n---\n## Core Principles\n1. **Strict reliance on the knowledge base**: answers must be **fully bounded** by the content retrieved from the knowledge base.\n2. **No fabrication**: do not generate, infer, or create information that is not explicitly present in the knowledge base.\n3. **Accuracy first**: prefer incompleteness over inaccurate content.\n4. **Output format**:\n   * Hierarchically clear modular structure\n   * Logical grouping according to the MECE principle\n   * Professionally presented formatting\n   * Step-by-step cognitive guidance\n   * Reasonable use of headings and dividers for clarity\n   * *Italicize* key parameters\n   * **Bold** critical information\n5. **LaTeX formula requirements**:\n   * Inline formulas: start and end with `$`\n   * Block formulas: start and end with `$$`, each `$$` on its own line\n   * Block formula content must comply with LaTeX math syntax\n   * Verify formula correctness\n---\n## Additional Notes (Interaction & Failure Strategy)\n* If the knowledge base does not cover critical facts: explicitly inform the user (with sample wording)\n* For time-sensitive issues: enforce time filtering in the search request, and indicate the latest retrieval date in the answer.\n* Language requirement: answer in the user\u2019s preferred language\n",
+                        "temperature": "0.1",
+                        "temperatureEnabled": true,
+                        "tools": [
+                            {
+                                "component_name": "Retrieval",
+                                "name": "Retrieval",
+                                "params": {
+                                    "cross_languages": [],
+                                    "description": "",
+                                    "empty_response": "",
+                                    "kb_ids": [],
+                                    "keywords_similarity_weight": 0.7,
+                                    "outputs": {
+                                        "formalized_content": {
+                                            "type": "string",
+                                            "value": ""
+                                        }
+                                    },
+                                    "rerank_id": "",
+                                    "similarity_threshold": 0.2,
+                                    "top_k": 1024,
+                                    "top_n": 8,
+                                    "use_kg": false
+                                }
+                            }
+                        ],
+                        "topPEnabled": false,
+                        "top_p": 0.75,
+                        "user_prompt": "",
+                        "visual_files_var": ""
+                    }
+                },
+                "upstream": [
+                    "begin"
+                ]
+            },
+            "Message:OrangeYearsShine": {
+                "downstream": [],
+                "obj": {
+                    "component_name": "Message",
+                    "params": {
+                        "content": [
+                            "{Agent:NewPumasLick@content}"
+                        ]
+                    }
+                },
+                "upstream": [
+                    "Agent:NewPumasLick"
+                ]
+            },
+            "begin": {
+                "downstream": [
+                    "Agent:NewPumasLick"
+                ],
+                "obj": {
+                    "component_name": "Begin",
+                    "params": {
+                        "enablePrologue": true,
+                        "inputs": {},
+                        "mode": "conversational",
+                        "prologue": "\u4f60\u597d\uff01 \u6211\u662f\u4f60\u7684\u52a9\u7406\uff0c\u6709\u4ec0\u4e48\u53ef\u4ee5\u5e2e\u5230\u4f60\u7684\u5417\uff1f"
+                    }
+                },
+                "upstream": []
+            }
+        },
+        "globals": {
+            "sys.conversation_turns": 0,
+            "sys.files": [],
+            "sys.query": "",
+            "sys.user_id": ""
+        },
+        "graph": {
+            "edges": [
+                {
+                    "data": {
+                        "isHovered": false
+                    },
+                    "id": "xy-edge__beginstart-Agent:NewPumasLickend",
+                    "source": "begin",
+                    "sourceHandle": "start",
+                    "target": "Agent:NewPumasLick",
+                    "targetHandle": "end"
+                },
+                {
+                    "data": {
+                        "isHovered": false
+                    },
+                    "id": "xy-edge__Agent:NewPumasLickstart-Message:OrangeYearsShineend",
+                    "markerEnd": "logo",
+                    "source": "Agent:NewPumasLick",
+                    "sourceHandle": "start",
+                    "style": {
+                        "stroke": "rgba(91, 93, 106, 1)",
+                        "strokeWidth": 1
+                    },
+                    "target": "Message:OrangeYearsShine",
+                    "targetHandle": "end",
+                    "type": "buttonEdge",
+                    "zIndex": 1001
+                },
+                {
+                    "data": {
+                        "isHovered": false
+                    },
+                    "id": "xy-edge__Agent:NewPumasLicktool-Tool:AllBirdsNailend",
+                    "selected": false,
+                    "source": "Agent:NewPumasLick",
+                    "sourceHandle": "tool",
+                    "target": "Tool:AllBirdsNail",
+                    "targetHandle": "end"
+                }
+            ],
+            "nodes": [
+                {
+                    "data": {
+                        "form": {
+                            "enablePrologue": true,
+                            "inputs": {},
+                            "mode": "conversational",
+                            "prologue": "\u4f60\u597d\uff01 \u6211\u662f\u4f60\u7684\u52a9\u7406\uff0c\u6709\u4ec0\u4e48\u53ef\u4ee5\u5e2e\u5230\u4f60\u7684\u5417\uff1f"
+                        },
+                        "label": "Begin",
+                        "name": "begin"
+                    },
+                    "dragging": false,
+                    "id": "begin",
+                    "measured": {
+                        "height": 48,
+                        "width": 200
+                    },
+                    "position": {
+                        "x": -9.569875358221438,
+                        "y": 205.84018385864917
+                    },
+                    "selected": false,
+                    "sourcePosition": "left",
+                    "targetPosition": "right",
+                    "type": "beginNode"
+                },
+                {
+                    "data": {
+                        "form": {
+                            "content": [
+                                "{Agent:NewPumasLick@content}"
+                            ]
+                        },
+                        "label": "Message",
+                        "name": "Response"
+                    },
+                    "dragging": false,
+                    "id": "Message:OrangeYearsShine",
+                    "measured": {
+                        "height": 56,
+                        "width": 200
+                    },
+                    "position": {
+                        "x": 734.4061285881053,
+                        "y": 199.9706031723009
+                    },
+                    "selected": false,
+                    "sourcePosition": "right",
+                    "targetPosition": "left",
+                    "type": "messageNode"
+                },
+                {
+                    "data": {
+                        "form": {
+                            "delay_after_error": 1,
+                            "description": "",
+                            "exception_comment": "",
+                            "exception_default_value": "",
+                            "exception_goto": [],
+                            "exception_method": null,
+                            "frequencyPenaltyEnabled": false,
+                            "frequency_penalty": 0.5,
+                            "llm_id": "qwen3-235b-a22b-instruct-2507@Tongyi-Qianwen",
+                            "maxTokensEnabled": true,
+                            "max_retries": 3,
+                            "max_rounds": 3,
+                            "max_tokens": 128000,
+                            "mcp": [],
+                            "message_history_window_size": 12,
+                            "outputs": {
+                                "content": {
+                                    "type": "string",
+                                    "value": ""
+                                }
+                            },
+                            "parameter": "Precise",
+                            "presencePenaltyEnabled": false,
+                            "presence_penalty": 0.5,
+                            "prompts": [
+                                {
+                                    "content": "# User Query\n {sys.query}",
+                                    "role": "user"
+                                }
+                            ],
+                            "sys_prompt": "## Role & Task\nYou are a **\u201cKnowledge Base Retrieval Q\\&A Agent\u201d** whose goal is to break down the user\u2019s question into retrievable subtasks, and then produce a multi-source-verified, structured, and actionable research report using the internal knowledge base.\n## Execution Framework (Detailed Steps & Key Points)\n1. **Assessment & Decomposition**\n   * Actions:\n     * Automatically extract: main topic, subtopics, entities (people/organizations/products/technologies), time window, geographic/business scope.\n     * Output as a list: N facts/data points that must be collected (*N* ranges from 5\u201320 depending on question complexity).\n2. **Query Type Determination (Rule-Based)**\n   * Example rules:\n     * If the question involves a single issue but requests \u201cmethod comparison/multiple explanations\u201d \u2192 use **depth-first**.\n     * If the question can naturally be split into \u22653 independent sub-questions \u2192 use **breadth-first**.\n     * If the question can be answered by a single fact/specification/definition \u2192 use **simple query**.\n3. **Research Plan Formulation**\n   * Depth-first: define 3\u20135 perspectives (methodology/stakeholders/time dimension/technical route, etc.), assign search keywords, target document types, and output format for each perspective.\n   * Breadth-first: list subtasks, prioritize them, and assign search terms.\n   * Simple query: directly provide the search sentence and required fields.\n4. **Retrieval Execution**\n   * After retrieval: perform coverage check (does it contain the key facts?) and quality check (source diversity, authority, latest update time).\n   * If standards are not met, automatically loop: rewrite queries (synonyms/cross-domain terms) and retry \u22643 times, or flag as requiring external search.\n5. **Integration & Reasoning**\n   * Build the answer using a **fact\u2013evidence\u2013reasoning** chain. For each conclusion, attach 1\u20132 strongest pieces of evidence.\n---\n## Quality Gate Checklist (Verify at Each Stage)\n* **Stage 1 (Decomposition)**:\n  * [ ] Key concepts and expected outputs identified\n  * [ ] Required facts/data points listed\n* **Stage 2 (Retrieval)**:\n  * [ ] Meets quality standards (see above)\n  * [ ] If not met: execute query iteration\n* **Stage 3 (Generation)**:\n  * [ ] Each conclusion has at least one direct evidence source\n  * [ ] State assumptions/uncertainties\n  * [ ] Provide next-step suggestions or experiment/retrieval plans\n  * [ ] Final length and depth match user expectations (comply with word count/format if specified)\n---\n## Core Principles\n1. **Strict reliance on the knowledge base**: answers must be **fully bounded** by the content retrieved from the knowledge base.\n2. **No fabrication**: do not generate, infer, or create information that is not explicitly present in the knowledge base.\n3. **Accuracy first**: prefer incompleteness over inaccurate content.\n4. **Output format**:\n   * Hierarchically clear modular structure\n   * Logical grouping according to the MECE principle\n   * Professionally presented formatting\n   * Step-by-step cognitive guidance\n   * Reasonable use of headings and dividers for clarity\n   * *Italicize* key parameters\n   * **Bold** critical information\n5. **LaTeX formula requirements**:\n   * Inline formulas: start and end with `$`\n   * Block formulas: start and end with `$$`, each `$$` on its own line\n   * Block formula content must comply with LaTeX math syntax\n   * Verify formula correctness\n---\n## Additional Notes (Interaction & Failure Strategy)\n* If the knowledge base does not cover critical facts: explicitly inform the user (with sample wording)\n* For time-sensitive issues: enforce time filtering in the search request, and indicate the latest retrieval date in the answer.\n* Language requirement: answer in the user\u2019s preferred language\n",
+                            "temperature": "0.1",
+                            "temperatureEnabled": true,
+                            "tools": [
+                                {
+                                    "component_name": "Retrieval",
+                                    "name": "Retrieval",
+                                    "params": {
+                                        "cross_languages": [],
+                                        "description": "",
+                                        "empty_response": "",
+                                        "kb_ids": [],
+                                        "keywords_similarity_weight": 0.7,
+                                        "outputs": {
+                                            "formalized_content": {
+                                                "type": "string",
+                                                "value": ""
+                                            }
+                                        },
+                                        "rerank_id": "",
+                                        "similarity_threshold": 0.2,
+                                        "top_k": 1024,
+                                        "top_n": 8,
+                                        "use_kg": false
+                                    }
+                                }
+                            ],
+                            "topPEnabled": false,
+                            "top_p": 0.75,
+                            "user_prompt": "",
+                            "visual_files_var": ""
+                        },
+                        "label": "Agent",
+                        "name": "Knowledge Base Agent"
+                    },
+                    "dragging": false,
+                    "id": "Agent:NewPumasLick",
+                    "measured": {
+                        "height": 84,
+                        "width": 200
+                    },
+                    "position": {
+                        "x": 347.00048227952215,
+                        "y": 186.49109364794631
+                    },
+                    "selected": false,
+                    "sourcePosition": "right",
+                    "targetPosition": "left",
+                    "type": "agentNode"
+                },
+                {
+                    "data": {
+                        "form": {
+                            "description": "This is an agent for a specific task.",
+                            "user_prompt": "This is the order you need to send to the agent."
+                        },
+                        "label": "Tool",
+                        "name": "flow.tool_10"
+                    },
+                    "dragging": false,
+                    "id": "Tool:AllBirdsNail",
+                    "measured": {
+                        "height": 48,
+                        "width": 200
+                    },
+                    "position": {
+                        "x": 220.24819746977118,
+                        "y": 403.31576836482583
+                    },
+                    "selected": false,
+                    "sourcePosition": "right",
+                    "targetPosition": "left",
+                    "type": "toolNode"
+                }
+            ]
+        },
+        "history": [],
+        "memory": [],
+        "messages": [],
+        "path": [],
+        "retrieval": []
+    },
+    "avatar": "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAADAAAAAwCAYAAABXAvmHAAAH0klEQVR4nO2ZC1BU1wGG/3uRp/IygG+DGK0GOjE1U6cxI4tT03Y0E+kENbaJbKpj60wzgNMwnTjuEtu0miGasY+0krI202kMVEnVxtoOLG00oVa0LajVBDcSEI0REFBgkZv/3GWXfdzdvctuHs7kmzmec9//d+45914XCXc4Xwjk1+59VJGGF7C5QAFSWBvgyWmWLl7IKiny6QNL173B5YjB84bOyrpKA4B1DLySdQpLKAiZGtZ7a/KMVoQJz6UfEZyhTWwaEBmssiLvCueu6BJg8EwFqGTTAC+uvNWC9w82sRWcux/JwaSHstjywcogRt4RG0KExwWG4QsVYCebKSwe3L5lR9OOWjyzfg2WL/0a1/jncO3b2FHxGnKeWYqo+Giu8UEMrWJKWBACPMY/DG+63txhvnKshUu+DF2/hayMDFRsL+VScDb++AVc6OjAuInxXPJl2tfnIikrzUyJMi7qQmLRhOEr2fOFbX/7P6STF7BqoWevfdij4NWGQfx+57OYO2sG1wSnsek8Nm15EU8sikF6ouelXz9ph7JwDqYt+5IIZaGEkauDIrH4wPBmhjexCSEws+VdVG1M4NIoj+2xYzBuJtavWcEl/VS8dggx/ZdQvcGzQwp+cxOXsu5RBQQMVkYJM4LA/Txh+ELFMWFVPARS5kFiabZdx8Olh7l17BzdvhzZmROhdJ3j6D/nIyBgOCMlLAgA9xmF4TMV4BSbrgnrLiBl5rOsRCRRbDUsBzQFiJjY91PCBj9w+yiP1lXWsTLAjc9YQGB9I8+Yx1oTiUWFvW9QgDo2PdASaDp/EQ8/sRnhcPTVcuTMncXwQQVESL9DidscaPW+QEtAICRu9PSxFTpJiePV8AI9AsTvXZBY/Pa+wJ9ApNApIILm8S5Y4QXXQwhYFH6csemDP4G3G5v579i5d04mknknQhDYS4HCrCVr/mC3D305KnbCEpvVIia5Onw6WaWw+KAl0Np+FUXbdiMcyoqfUoeRHoFrJ1uRtnBG1/9Mf/3LtElp+VwF2wcd7woJib1vUPwMH4GWQCQJJtBa/V9cPmFD8uQUpMdNGDhY8bNYrobh8acHu270/l0ImJWRt64Wn6WACN9z5gq2lXwPW8pfweT0icP/fH23vO9QLYq3/QKyLBmFQI3CUcT9NdESEEPItKsSN3r7MBaSJoxHWZERM6ZmMLy2gDP8/pd/og418dTL37hFSUpMUC5f+UiWZcnY9s5+ixCwUiCXx2iiJdDNx6f4pgkH8Q3lbxK7h8+enoHha1cRNdMp8axiHxo6+/5bVdk8DSROYIW1X7QEIom3wHD3gEf4vu1bVYEJZeWQ0zJQvmcfyiv2QZak6raG/QWfK4Ez9mTc5v8xPMJfuojoxXmIX/9DOMe+FCWbcHu4BJJ0YEwCx0824bFNW9HesB+CqYu+jepfPYcHF+aoPXS8sQl/+vU2bgmOU2C+qRc9/YrrPPbGBtzavd0nvCxLxui4pJrBm911PFwak4CYA80cj+JCAiGUzYkmxrSY4N2c3GLi6UEIFL/wRxxqkhmHnTEpDQcrfq6ea+hcE8bNy3GFzyq4H22HW1Kd4WMSkg1jmsSRpKj0Rzhy4gNUv/y8Gjrv8SJK3OWScA+fMn/ysVPPvTmeh6nh1TcxBUJ+jEaKYr7N36x7h+Edj0pB6+WrLokn87+BrTt/p4ZPzZ6MM7/8R2//h33vOcNzdwgBMwVMbGvySQmo4a0NqOZccU7YmGXLEfPQUlUid/XT6B8YdIU/99vjsPcOdEhDsfOd4QVCwKB8yp8SWuG1njbTl83DpMWz1PCKAswuWPDI0e8WebyAJBbxNdrF7cls+hBpAb3h3XtehL/3+4u7D35rQwpP4YFTwMJ91rHpQyQFQgmf9sAMNL9Ur4afv/FBjIuPVj+n4YVTwMD96tj0IVICoYYXv/q1VJ1Sl8UveQyaRwErvOB6B5SwKhqP00gI6A0vhsycJ7/KIzxhyHqGN0ADbnNAAYOicRfCFdAb/p50Gbfuc/wy5w1D5lOghk0fuG0USlgVr7sQjoDe8C8WxKGKPy2KjzlvAQb02/sCbh+FApngX1QUtyeSuwDi0hxFByV7L+LIf3r5kvpp4PBr07Hqvn71Y85bgOG6WS2ggA1+4D6eUKKQApVsqngI6KSkqh9HzsoM/3zg8Oz5VQ9E8wjf30YFDGdkeAsCwH18oYRZGXk7C4HuYxcwe6rjQsFovzaEvoFxqNkTOPzMjGikJso8wsF77XYkLx6dAwxWxvBmBIH7aUMJi8J3w0DnTVz7dyvX6KPzVBt+kL8cmzesRq9ps2Z48bRJmOIapS7E4zM2lXNt5CcU6ID7+ocSZkqY2NRN6ysnsHbJEpR8ZwV6t5Yg+iuLELf2KVd48VwXQf3BQGUMb4ZOuH9gKFEIYJfiNrEDcXZHHV4q3YRv5i7ikgM94RlETNgihrcgBHhccCiRCf7VhBK5rAPyr9I/Y/WKPEyfksH/9NjQ2dODhsYzwcLXsypkeBtCRGLRDUUMAMyKHxEx4dtrzyP97nQMygripiQiKi4aSbPvQmKW7+OXF69ntYvBa1iPCYklZEZECsGm4ja0Ops7EJsaj4SprlU+8IJiqIjAFga3Ikx4vvAYkTGALxyWFArlsnbBC9Sz6mI5zWKNRGh3JJY7mjte4GOz+r4tkRbxQQAAAABJRU5ErkJggg=="
+}

agent/templates/sql_assistant.json

@@ -206,7 +206,7 @@
                             "enablePrologue": true,
                             "inputs": {},
                             "mode": "conversational",
-                            "prologue": "Hi! I'm your SQL assistant, what can I do for you?"
+                            "prologue": "Hi! I'm your SQL assistant. What can I do for you?"
                         }
                     },
                     "upstream": []
@@ -319,7 +319,7 @@
                                 "enablePrologue": true,
                                 "inputs": {},
                                 "mode": "conversational",
-                                "prologue": "Hi! I'm your SQL assistant, what can I do for you?"
+                                "prologue": "Hi! I'm your SQL assistant. What can I do for you?"
                             },
                             "label": "Begin",
                             "name": "begin"

agent/tools/base.py

@@ -24,6 +24,7 @@ from api.utils import hash_str2int
 from rag.llm.chat_model import ToolCallSession
 from rag.prompts.prompts import kb_prompt
 from rag.utils.mcp_tool_call_conn import MCPToolCallSession
+from timeit import default_timer as timer
 
 
 class ToolParameter(TypedDict):
@@ -49,12 +50,13 @@ class LLMToolPluginCallSession(ToolCallSession):
 
     def tool_call(self, name: str, arguments: dict[str, Any]) -> Any:
         assert name in self.tools_map, f"LLM tool {name} does not exist"
+        st = timer()
         if isinstance(self.tools_map[name], MCPToolCallSession):
             resp = self.tools_map[name].tool_call(name, arguments, 60)
         else:
             resp = self.tools_map[name].invoke(**arguments)
 
-        self.callback(name, arguments, resp)
+        self.callback(name, arguments, resp, elapsed_time=timer()-st)
         return resp
 
     def get_tool_obj(self, name):

agent/tools/code_exec.py

@@ -17,7 +17,7 @@ import base64
 import logging
 import os
 from abc import ABC
-from enum import StrEnum
+from strenum import StrEnum
 from typing import Optional
 from pydantic import BaseModel, Field, field_validator
 from agent.tools.base import ToolParamBase, ToolBase, ToolMeta
@@ -67,9 +67,17 @@ class CodeExecParam(ToolParamBase):
             "description": """
 This tool has a sandbox that can execute code written in 'Python'/'Javascript'. It recieves a piece of code and return a Json string.
 Here's a code example for Python(`main` function MUST be included):
-def main(arg1: str, arg2: str) -> dict:
+def main() -> dict:
+    \"\"\"
+    Generate Fibonacci numbers within 100.
+    \"\"\"
+    def fibonacci_recursive(n):
+        if n <= 1:
+            return n
+        else:
+            return fibonacci_recursive(n-1) + fibonacci_recursive(n-2)
     return {
-        "result": arg1 + arg2,
+        "result": fibonacci_recursive(100),
     }
     
 Here's a code example for Javascript(`main` function MUST be included and exported):

agent/tools/exesql.py

@@ -14,6 +14,7 @@
 #  limitations under the License.
 #
 import os
+import re
 from abc import ABC
 import pandas as pd
 import pymysql
@@ -78,6 +79,17 @@ class ExeSQL(ToolBase, ABC):
 
     @timeout(os.environ.get("COMPONENT_EXEC_TIMEOUT", 60))
     def _invoke(self, **kwargs):
+
+        def convert_decimals(obj):
+            from decimal import Decimal
+            if isinstance(obj, Decimal):
+                return float(obj)  # 或 str(obj)
+            elif isinstance(obj, dict):
+                return {k: convert_decimals(v) for k, v in obj.items()}
+            elif isinstance(obj, list):
+                return [convert_decimals(item) for item in obj]
+            return obj
+
         sql = kwargs.get("sql")
         if not sql:
             raise Exception("SQL for `ExeSQL` MUST not be empty.")
@@ -109,7 +121,7 @@ class ExeSQL(ToolBase, ABC):
             single_sql = single_sql.replace('```','')
             if not single_sql:
                 continue
-
+            single_sql = re.sub(r"\[ID:[0-9]+\]", "", single_sql)
             cursor.execute(single_sql)
             if cursor.rowcount == 0:
                 sql_res.append({"content": "No record in the database!"})
@@ -121,7 +133,11 @@ class ExeSQL(ToolBase, ABC):
                 single_res = pd.DataFrame([i for i in cursor.fetchmany(self._param.max_records)])
                 single_res.columns = [i[0] for i in cursor.description]
 
-            sql_res.append(single_res.to_dict(orient='records'))
+            for col in single_res.columns:
+                if pd.api.types.is_datetime64_any_dtype(single_res[col]):
+                    single_res[col] = single_res[col].dt.strftime('%Y-%m-%d')
+
+            sql_res.append(convert_decimals(single_res.to_dict(orient='records')))
             formalized_content.append(single_res.to_markdown(index=False, floatfmt=".6f"))
 
         self.set_output("json", sql_res)
@@ -129,4 +145,4 @@ class ExeSQL(ToolBase, ABC):
         return self.output("formalized_content")
 
     def thoughts(self) -> str:
-        return "Query sent—waiting for the data."
+        return "Query sent—waiting for the data."

agent/tools/retrieval.py

@@ -86,10 +86,16 @@ class Retrieval(ToolBase, ABC):
                 kb_ids.append(id)
                 continue
             kb_nm = self._canvas.get_variable_value(id)
-            e, kb = KnowledgebaseService.get_by_name(kb_nm, self._canvas._tenant_id)
-            if not e:
-                raise Exception(f"Dataset({kb_nm}) does not exist.")
-            kb_ids.append(kb.id)
+            # if kb_nm is a list
+            kb_nm_list = kb_nm if isinstance(kb_nm, list) else [kb_nm]
+            for nm_or_id in kb_nm_list:
+                e, kb = KnowledgebaseService.get_by_name(nm_or_id,
+                                                         self._canvas._tenant_id)
+                if not e:
+                    e, kb = KnowledgebaseService.get_by_id(nm_or_id)
+                    if not e:
+                        raise Exception(f"Dataset({nm_or_id}) does not exist.")
+                kb_ids.append(kb.id)
 
         filtered_kb_ids: list[str] = list(set([kb_id for kb_id in kb_ids if kb_id]))
 
@@ -108,7 +114,9 @@ class Retrieval(ToolBase, ABC):
         if self._param.rerank_id:
             rerank_mdl = LLMBundle(kbs[0].tenant_id, LLMType.RERANK, self._param.rerank_id)
 
-        query = kwargs["query"]
+        vars = self.get_input_elements_from_text(kwargs["query"])
+        vars = {k:o["value"] for k,o in vars.items()}
+        query = self.string_format(kwargs["query"], vars)
         if self._param.cross_languages:
             query = cross_languages(kbs[0].tenant_id, None, query, self._param.cross_languages)
 

api/apps/__init__.py

@@ -29,6 +29,7 @@ from api.db.db_models import close_connection
 from api.db.services import UserService
 from api.utils import CustomJSONEncoder, commands
 
+from flask_mail import Mail
 from flask_session import Session
 from flask_login import LoginManager
 from api import settings
@@ -40,6 +41,7 @@ __all__ = ["app"]
 Request.json = property(lambda self: self.get_json(force=True, silent=True))
 
 app = Flask(__name__)
+smtp_mail_server = Mail()
 
 # Add this at the beginning of your file to configure Swagger UI
 swagger_config = {
@@ -146,16 +148,16 @@ def load_user(web_request):
     if authorization:
         try:
             access_token = str(jwt.loads(authorization))
-            
+
             if not access_token or not access_token.strip():
                 logging.warning("Authentication attempt with empty access token")
                 return None
-            
+
             # Access tokens should be UUIDs (32 hex characters)
             if len(access_token.strip()) < 32:
                 logging.warning(f"Authentication attempt with invalid token format: {len(access_token)} chars")
                 return None
-            
+
             user = UserService.query(
                 access_token=access_token, status=StatusEnum.VALID.value
             )

api/apps/canvas_app.py

@@ -74,11 +74,11 @@ def rm():
 @login_required
 def save():
     req = request.json
-    req["user_id"] = current_user.id
     if not isinstance(req["dsl"], str):
         req["dsl"] = json.dumps(req["dsl"], ensure_ascii=False)
     req["dsl"] = json.loads(req["dsl"])
     if "id" not in req:
+        req["user_id"] = current_user.id
         if UserCanvasService.query(user_id=current_user.id, title=req["title"].strip()):
             return get_data_error_result(message=f"{req['title'].strip()} already exists.")
         req["id"] = get_uuid()
@@ -90,7 +90,7 @@ def save():
                 data=False, message='Only owner of canvas authorized for this operation.',
                 code=RetCode.OPERATING_ERROR)
         UserCanvasService.update_by_id(req["id"], req)
-    # save version    
+    # save version
     UserCanvasVersionService.insert( user_canvas_id=req["id"], dsl=req["dsl"], title="{0}_{1}".format(req["title"], time.strftime("%Y_%m_%d_%H_%M_%S")))
     UserCanvasVersionService.delete_all_versions(req["id"])
     return get_json_result(data=req)
@@ -347,7 +347,7 @@ def test_db_connect():
         if req["db_type"] != 'mssql':
             db.connect()
         db.close()
-        
+
         return get_json_result(data="Database Connection Successful!")
     except Exception as e:
         return server_error_response(e)
@@ -369,7 +369,7 @@ def getlistversion(canvas_id):
 @login_required
 def getversion( version_id):
     try:
-      
+
         e, version = UserCanvasVersionService.get_by_id(version_id)
         if version:
             return get_json_result(data=version.to_dict())
@@ -379,7 +379,7 @@ def getversion( version_id):
 
 @manager.route('/listteam', methods=['GET'])  # noqa: F821
 @login_required
-def list_kbs():
+def list_canvas():
     keywords = request.args.get("keywords", "")
     page_number = int(request.args.get("page", 1))
     items_per_page = int(request.args.get("page_size", 150))
@@ -387,10 +387,10 @@ def list_kbs():
     desc = request.args.get("desc", True)
     try:
         tenants = TenantService.get_joined_tenants_by_user_id(current_user.id)
-        kbs, total = UserCanvasService.get_by_tenant_ids(
+        canvas, total = UserCanvasService.get_by_tenant_ids(
             [m["tenant_id"] for m in tenants], current_user.id, page_number,
             items_per_page, orderby, desc, keywords)
-        return get_json_result(data={"kbs": kbs, "total": total})
+        return get_json_result(data={"canvas": canvas, "total": total})
     except Exception as e:
         return server_error_response(e)
 

api/apps/conversation_app.py

@@ -28,11 +28,14 @@ from api.db.db_models import APIToken
 from api.db.services.conversation_service import ConversationService, structure_answer
 from api.db.services.dialog_service import DialogService, ask, chat
 from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.db.services.llm_service import LLMBundle, TenantService
-from api.db.services.user_service import UserTenantService
+from api.db.services.llm_service import LLMBundle
+from api.db.services.search_service import SearchService
+from api.db.services.tenant_llm_service import TenantLLMService
+from api.db.services.user_service import TenantService, UserTenantService
 from api.utils.api_utils import get_data_error_result, get_json_result, server_error_response, validate_request
 from graphrag.general.mind_map_extractor import MindMapExtractor
 from rag.app.tag import label_question
+from rag.prompts.prompt_template import load_prompt
 from rag.prompts.prompts import chunks_format
 
 
@@ -66,8 +69,14 @@ 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,
-                "reference":[{}],}
+        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:
@@ -174,6 +183,21 @@ def completion():
             continue
         msg.append(m)
     message_id = msg[-1].get("id")
+    chat_model_id = req.get("llm_id", "")
+    req.pop("llm_id", None)
+
+    chat_model_config = {}
+    for model_config in [
+        "temperature",
+        "top_p",
+        "frequency_penalty",
+        "presence_penalty",
+        "max_tokens",
+    ]:
+        config = req.get(model_config)
+        if config:
+            chat_model_config[model_config] = config
+
     try:
         e, conv = ConversationService.get_by_id(req["conversation_id"])
         if not e:
@@ -187,23 +211,26 @@ def completion():
 
         if not conv.reference:
             conv.reference = []
-        else:
-            for ref in conv.reference:
-                if isinstance(ref, list):
-                    continue
-                ref["chunks"] = chunks_format(ref)
-
-        if not conv.reference:
-            conv.reference = []
+        conv.reference = [r for r in conv.reference if r]
         conv.reference.append({"chunks": [], "doc_aggs": []})
 
+        if chat_model_id:
+            if not TenantLLMService.get_api_key(tenant_id=dia.tenant_id, model_name=chat_model_id):
+                req.pop("chat_model_id", None)
+                req.pop("chat_model_config", None)
+                return get_data_error_result(message=f"Cannot use specified model {chat_model_id}.")
+            dia.llm_id = chat_model_id
+            dia.llm_setting = chat_model_config
+
+        is_embedded = bool(chat_model_id)
         def stream():
             nonlocal dia, msg, req, conv
             try:
                 for ans in chat(dia, msg, True, **req):
                     ans = structure_answer(conv, ans, message_id, conv.id)
                     yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
-                ConversationService.update_by_id(conv.id, conv.to_dict())
+                if not is_embedded:
+                    ConversationService.update_by_id(conv.id, conv.to_dict())
             except Exception as e:
                 traceback.print_exc()
                 yield "data:" + json.dumps({"code": 500, "message": str(e), "data": {"answer": "**ERROR**: " + str(e), "reference": []}}, ensure_ascii=False) + "\n\n"
@@ -221,7 +248,8 @@ def completion():
             answer = None
             for ans in chat(dia, msg, **req):
                 answer = structure_answer(conv, ans, message_id, conv.id)
-                ConversationService.update_by_id(conv.id, conv.to_dict())
+                if not is_embedded:
+                    ConversationService.update_by_id(conv.id, conv.to_dict())
                 break
             return get_json_result(data=answer)
     except Exception as e:
@@ -317,10 +345,18 @@ def ask_about():
     req = request.json
     uid = current_user.id
 
+    search_id = req.get("search_id", "")
+    search_app = None
+    search_config = {}
+    if search_id:
+        search_app = SearchService.get_detail(search_id)
+    if search_app:
+        search_config = search_app.get("search_config", {})
+
     def stream():
         nonlocal req, uid
         try:
-            for ans in ask(req["question"], req["kb_ids"], uid):
+            for ans in ask(req["question"], req["kb_ids"], uid, search_config=search_config):
                 yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
         except Exception as e:
             yield "data:" + json.dumps({"code": 500, "message": str(e), "data": {"answer": "**ERROR**: " + str(e), "reference": []}}, ensure_ascii=False) + "\n\n"
@@ -339,15 +375,68 @@ def ask_about():
 @validate_request("question", "kb_ids")
 def mindmap():
     req = request.json
+
+    search_id = req.get("search_id", "")
+    search_app = None
+    search_config = {}
+    if search_id:
+        search_app = SearchService.get_detail(search_id)
+    if search_app:
+        search_config = search_app.get("search_config", {})
+
     kb_ids = req["kb_ids"]
+    if search_config.get("kb_ids", []):
+        kb_ids = search_config.get("kb_ids", [])
     e, kb = KnowledgebaseService.get_by_id(kb_ids[0])
     if not e:
         return get_data_error_result(message="Knowledgebase not found!")
 
-    embd_mdl = LLMBundle(kb.tenant_id, LLMType.EMBEDDING, llm_name=kb.embd_id)
-    chat_mdl = LLMBundle(current_user.id, LLMType.CHAT)
+    chat_id = ""
+    similarity_threshold = 0.3,
+    vector_similarity_weight = 0.3,
+    top = 1024,
+    doc_ids = []
+    rerank_id = ""
+    rerank_mdl = None
+
+    if search_config:
+        if search_config.get("chat_id", ""):
+            chat_id = search_config.get("chat_id", "")
+        if search_config.get("similarity_threshold", 0.2):
+            similarity_threshold = search_config.get("similarity_threshold", 0.2)
+        if search_config.get("vector_similarity_weight", 0.3):
+            vector_similarity_weight = search_config.get("vector_similarity_weight", 0.3)
+        if search_config.get("top_k", 1024):
+            top = search_config.get("top_k", 1024)
+        if search_config.get("doc_ids", []):
+            doc_ids = search_config.get("doc_ids", [])
+        if search_config.get("rerank_id", ""):
+            rerank_id = search_config.get("rerank_id", "")
+
+    tenant_id = kb.tenant_id
+    if search_app and search_app.get("tenant_id", ""):
+        tenant_id = search_app.get("tenant_id", "")
+
+    embd_mdl = LLMBundle(tenant_id, LLMType.EMBEDDING, llm_name=kb.embd_id)
+    chat_mdl = LLMBundle(tenant_id, LLMType.CHAT, llm_name=chat_id)
+    if rerank_id:
+        rerank_mdl = LLMBundle(tenant_id, LLMType.RERANK, rerank_id)
     question = req["question"]
-    ranks = settings.retrievaler.retrieval(question, embd_mdl, kb.tenant_id, kb_ids, 1, 12, 0.3, 0.3, aggs=False, rank_feature=label_question(question, [kb]))
+    ranks = settings.retrievaler.retrieval(
+        question=question,
+        embd_mdl=embd_mdl,
+        tenant_ids=tenant_id,
+        kb_ids=kb_ids,
+        page=1,
+        page_size=12,
+        similarity_threshold=similarity_threshold,
+        vector_similarity_weight=vector_similarity_weight,
+        top=top,
+        doc_ids=doc_ids,
+        aggs=False,
+        rerank_mdl=rerank_mdl,
+        rank_feature=label_question(question, [kb]),
+    )
     mindmap = MindMapExtractor(chat_mdl)
     mind_map = trio.run(mindmap, [c["content_with_weight"] for c in ranks["chunks"]])
     mind_map = mind_map.output
@@ -361,41 +450,20 @@ def mindmap():
 @validate_request("question")
 def related_questions():
     req = request.json
+
+    search_id = req.get("search_id", "")
+    search_config = {}
+    if search_id:
+        if search_app := SearchService.get_detail(search_id):
+            search_config = search_app.get("search_config", {})
+
     question = req["question"]
-    chat_mdl = LLMBundle(current_user.id, LLMType.CHAT)
-    prompt = """
-Role: You are an AI language model assistant tasked with generating 5-10 related questions based on a user’s original query. These questions should help expand the search query scope and improve search relevance.
 
-Instructions:
-	Input: You are provided with a user’s question.
-	Output: Generate 5-10 alternative questions that are related to the original user question. These alternatives should help retrieve a broader range of relevant documents from a vector database.
-	Context: Focus on rephrasing the original question in different ways, making sure the alternative questions are diverse but still connected to the topic of the original query. Do not create overly obscure, irrelevant, or unrelated questions.
-	Fallback: If you cannot generate any relevant alternatives, do not return any questions.
-	Guidance:
-	1. Each alternative should be unique but still relevant to the original query.
-	2. Keep the phrasing clear, concise, and easy to understand.
-	3. Avoid overly technical jargon or specialized terms unless directly relevant.
-	4. Ensure that each question contributes towards improving search results by broadening the search angle, not narrowing it.
+    chat_id = search_config.get("chat_id", "")
+    chat_mdl = LLMBundle(current_user.id, LLMType.CHAT, chat_id)
 
-Example:
-Original Question: What are the benefits of electric vehicles?
-
-Alternative Questions:
-	1. How do electric vehicles impact the environment?
-	2. What are the advantages of owning an electric car?
-	3. What is the cost-effectiveness of electric vehicles?
-	4. How do electric vehicles compare to traditional cars in terms of fuel efficiency?
-	5. What are the environmental benefits of switching to electric cars?
-	6. How do electric vehicles help reduce carbon emissions?
-	7. Why are electric vehicles becoming more popular?
-	8. What are the long-term savings of using electric vehicles?
-	9. How do electric vehicles contribute to sustainability?
-	10. What are the key benefits of electric vehicles for consumers?
-
-Reason:
-	Rephrasing the original query into multiple alternative questions helps the user explore different aspects of their search topic, improving the quality of search results.
-	These questions guide the search engine to provide a more comprehensive set of relevant documents.
-"""
+    gen_conf = search_config.get("llm_setting", {"temperature": 0.9})
+    prompt = load_prompt("related_question")
     ans = chat_mdl.chat(
         prompt,
         [
@@ -407,6 +475,6 @@ Related search terms:
     """,
             }
         ],
-        {"temperature": 0.9},
+        gen_conf,
     )
     return get_json_result(data=[re.sub(r"^[0-9]\. ", "", a) for a in ans.split("\n") if re.match(r"^[0-9]\. ", a)])

api/apps/dialog_app.py

@@ -18,7 +18,7 @@ from flask import request
 from flask_login import login_required, current_user
 from api.db.services.dialog_service import DialogService
 from api.db import StatusEnum
-from api.db.services.llm_service import TenantLLMService
+from api.db.services.tenant_llm_service import TenantLLMService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.user_service import TenantService, UserTenantService
 from api import settings
@@ -51,6 +51,7 @@ def set_dialog():
     similarity_threshold = req.get("similarity_threshold", 0.1)
     vector_similarity_weight = req.get("vector_similarity_weight", 0.3)
     llm_setting = req.get("llm_setting", {})
+    meta_data_filter = req.get("meta_data_filter", {})
     prompt_config = req["prompt_config"]
 
     if not is_create:
@@ -85,6 +86,7 @@ def set_dialog():
                 "llm_id": llm_id,
                 "llm_setting": llm_setting,
                 "prompt_config": prompt_config,
+                "meta_data_filter": meta_data_filter,
                 "top_n": top_n,
                 "top_k": top_k,
                 "rerank_id": rerank_id,

api/apps/document_app.py

@@ -681,6 +681,11 @@ def set_meta():
         return get_json_result(data=False, message="No authorization.", code=settings.RetCode.AUTHENTICATION_ERROR)
     try:
         meta = json.loads(req["meta"])
+        if not isinstance(meta, dict):
+            return get_json_result(data=False, message="Only dictionary type supported.", code=settings.RetCode.ARGUMENT_ERROR)
+        for k,v in meta.items():
+            if not isinstance(v, str) and not isinstance(v, int) and not isinstance(v, float):
+                return get_json_result(data=False, message=f"The type is not supported: {v}", code=settings.RetCode.ARGUMENT_ERROR)
     except Exception as e:
         return get_json_result(data=False, message=f"Json syntax error: {e}", code=settings.RetCode.ARGUMENT_ERROR)
     if not isinstance(meta, dict):

api/apps/kb_app.py

@@ -351,6 +351,7 @@ def knowledge_graph(kb_id):
             obj["graph"]["edges"] = sorted(filtered_edges, key=lambda x: x.get("weight", 0), reverse=True)[:128]
     return get_json_result(data=obj)
 
+
 @manager.route('/<kb_id>/knowledge_graph', methods=['DELETE'])  # noqa: F821
 @login_required
 def delete_knowledge_graph(kb_id):
@@ -364,3 +365,17 @@ def delete_knowledge_graph(kb_id):
     settings.docStoreConn.delete({"knowledge_graph_kwd": ["graph", "subgraph", "entity", "relation"]}, search.index_name(kb.tenant_id), kb_id)
 
     return get_json_result(data=True)
+
+
+@manager.route("/get_meta", methods=["GET"])  # noqa: F821
+@login_required
+def get_meta():
+    kb_ids = request.args.get("kb_ids", "").split(",")
+    for kb_id in kb_ids:
+        if not KnowledgebaseService.accessible(kb_id, current_user.id):
+            return get_json_result(
+                data=False,
+                message='No authorization.',
+                code=settings.RetCode.AUTHENTICATION_ERROR
+            )
+    return get_json_result(data=DocumentService.get_meta_by_kbs(kb_ids))

api/apps/llm_app.py

@@ -17,7 +17,8 @@ import logging
 import json
 from flask import request
 from flask_login import login_required, current_user
-from api.db.services.llm_service import LLMFactoriesService, TenantLLMService, LLMService
+from api.db.services.tenant_llm_service import LLMFactoriesService, TenantLLMService
+from api.db.services.llm_service import LLMService
 from api import settings
 from api.utils.api_utils import server_error_response, get_data_error_result, validate_request
 from api.db import StatusEnum, LLMType
@@ -57,6 +58,7 @@ def set_api_key():
     # test if api key works
     chat_passed, embd_passed, rerank_passed = False, False, False
     factory = req["llm_factory"]
+    extra = {"provider": factory}
     msg = ""
     for llm in LLMService.query(fid=factory):
         if not embd_passed and llm.model_type == LLMType.EMBEDDING.value:
@@ -73,7 +75,7 @@ def set_api_key():
         elif not chat_passed and llm.model_type == LLMType.CHAT.value:
             assert factory in ChatModel, f"Chat model from {factory} is not supported yet."
             mdl = ChatModel[factory](
-                req["api_key"], llm.llm_name, base_url=req.get("base_url"))
+                req["api_key"], llm.llm_name, base_url=req.get("base_url"), **extra)
             try:
                 m, tc = mdl.chat(None, [{"role": "user", "content": "Hello! How are you doing!"}],
                                  {"temperature": 0.9, 'max_tokens': 50})
@@ -204,6 +206,7 @@ def add_llm():
 
     msg = ""
     mdl_nm = llm["llm_name"].split("___")[0]
+    extra = {"provider": factory}
     if llm["model_type"] == LLMType.EMBEDDING.value:
         assert factory in EmbeddingModel, f"Embedding model from {factory} is not supported yet."
         mdl = EmbeddingModel[factory](
@@ -221,7 +224,8 @@ def add_llm():
         mdl = ChatModel[factory](
             key=llm['api_key'],
             model_name=mdl_nm,
-            base_url=llm["api_base"]
+            base_url=llm["api_base"],
+            **extra,
         )
         try:
             m, tc = mdl.chat(None, [{"role": "user", "content": "Hello! How are you doing!"}], {
@@ -312,12 +316,12 @@ def delete_factory():
 def my_llms():
     try:
         include_details = request.args.get('include_details', 'false').lower() == 'true'
-        
+
         if include_details:
             res = {}
             objs = TenantLLMService.query(tenant_id=current_user.id)
             factories = LLMFactoriesService.query(status=StatusEnum.VALID.value)
-            
+
             for o in objs:
                 o_dict = o.to_dict()
                 factory_tags = None
@@ -325,13 +329,13 @@ def my_llms():
                     if f.name == o_dict["llm_factory"]:
                         factory_tags = f.tags
                         break
-                        
+
                 if o_dict["llm_factory"] not in res:
                     res[o_dict["llm_factory"]] = {
                         "tags": factory_tags,
                         "llm": []
                     }
-                
+
                 res[o_dict["llm_factory"]]["llm"].append({
                     "type": o_dict["model_type"],
                     "name": o_dict["llm_name"],
@@ -352,7 +356,7 @@ def my_llms():
                     "name": o["llm_name"],
                     "used_token": o["used_tokens"]
                 })
-        
+
         return get_json_result(data=res)
     except Exception as e:
         return server_error_response(e)

api/apps/sdk/chat.py

@@ -21,7 +21,7 @@ from api import settings
 from api.db import StatusEnum
 from api.db.services.dialog_service import DialogService
 from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.db.services.llm_service import TenantLLMService
+from api.db.services.tenant_llm_service import TenantLLMService
 from api.db.services.user_service import TenantService
 from api.utils import get_uuid
 from api.utils.api_utils import check_duplicate_ids, get_error_data_result, get_result, token_required
@@ -99,7 +99,7 @@ def create(tenant_id):
       Here is the knowledge base:
       {knowledge}
       The above is the knowledge base.""",
-        "prologue": "Hi! I'm your assistant, what can I do for you?",
+        "prologue": "Hi! I'm your assistant. What can I do for you?",
         "parameters": [{"key": "knowledge", "optional": False}],
         "empty_response": "Sorry! No relevant content was found in the knowledge base!",
         "quote": True,
@@ -139,7 +139,7 @@ def create(tenant_id):
     res["llm"] = res.pop("llm_setting")
     res["llm"]["model_name"] = res.pop("llm_id")
     del res["kb_ids"]
-    res["dataset_ids"] = req["dataset_ids"]
+    res["dataset_ids"] = req.get("dataset_ids", [])
     res["avatar"] = res.pop("icon")
     return get_result(data=res)
 

api/apps/sdk/doc.py

@@ -32,7 +32,8 @@ from api.db.services.document_service import DocumentService
 from api.db.services.file2document_service import File2DocumentService
 from api.db.services.file_service import FileService
 from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.db.services.llm_service import LLMBundle, TenantLLMService
+from api.db.services.llm_service import LLMBundle
+from api.db.services.tenant_llm_service import TenantLLMService
 from api.db.services.task_service import TaskService, queue_tasks
 from api.utils.api_utils import check_duplicate_ids, construct_json_result, get_error_data_result, get_parser_config, get_result, server_error_response, token_required
 from rag.app.qa import beAdoc, rmPrefix

api/apps/sdk/session.py

@@ -16,11 +16,11 @@
 import json
 import re
 import time
-
 import tiktoken
 from flask import Response, jsonify, request
-
+import trio
 from agent.canvas import Canvas
+from api import settings
 from api.db import LLMType, StatusEnum
 from api.db.db_models import APIToken
 from api.db.services.api_service import API4ConversationService
@@ -29,12 +29,17 @@ from api.db.services.canvas_service import completion as agent_completion
 from api.db.services.conversation_service import ConversationService, iframe_completion
 from api.db.services.conversation_service import completion as rag_completion
 from api.db.services.dialog_service import DialogService, ask, chat
-from api.db.services.file_service import FileService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.llm_service import LLMBundle
+from api.db.services.search_service import SearchService
+from api.db.services.user_service import UserTenantService
 from api.utils import get_uuid
-from api.utils.api_utils import check_duplicate_ids, get_data_openai, get_error_data_result, get_result, token_required, validate_request
+from api.utils.api_utils import check_duplicate_ids, get_data_openai, get_error_data_result, get_json_result, get_result, server_error_response, token_required, validate_request
+from graphrag.general.mind_map_extractor import MindMapExtractor
+from rag.app.tag import label_question
 from rag.prompts import chunks_format
+from rag.prompts.prompt_template import load_prompt
+from rag.prompts.prompts import cross_languages, keyword_extraction
 
 
 @manager.route("/chats/<chat_id>/sessions", methods=["POST"])  # noqa: F821
@@ -69,11 +74,7 @@ def create(tenant_id, chat_id):
 @manager.route("/agents/<agent_id>/sessions", methods=["POST"])  # noqa: F821
 @token_required
 def create_agent_session(tenant_id, agent_id):
-    req = request.json
-    if not request.is_json:
-        req = request.form
-    files = request.files
-    user_id = request.args.get("user_id", "")
+    user_id = request.args.get("user_id", tenant_id)
     e, cvs = UserCanvasService.get_by_id(agent_id)
     if not e:
         return get_error_data_result("Agent not found.")
@@ -82,46 +83,21 @@ def create_agent_session(tenant_id, agent_id):
     if not isinstance(cvs.dsl, str):
         cvs.dsl = json.dumps(cvs.dsl, ensure_ascii=False)
 
-    canvas = Canvas(cvs.dsl, tenant_id)
+    session_id=get_uuid()
+    canvas = Canvas(cvs.dsl, tenant_id, agent_id)
     canvas.reset()
-    query = canvas.get_preset_param()
-    if query:
-        for ele in query:
-            if not ele["optional"]:
-                if ele["type"] == "file":
-                    if files is None or not files.get(ele["key"]):
-                        return get_error_data_result(f"`{ele['key']}` with type `{ele['type']}` is required")
-                    upload_file = files.get(ele["key"])
-                    file_content = FileService.parse_docs([upload_file], user_id)
-                    file_name = upload_file.filename
-                    ele["value"] = file_name + "\n" + file_content
-                else:
-                    if req is None or not req.get(ele["key"]):
-                        return get_error_data_result(f"`{ele['key']}` with type `{ele['type']}` is required")
-                    ele["value"] = req[ele["key"]]
-            else:
-                if ele["type"] == "file":
-                    if files is not None and files.get(ele["key"]):
-                        upload_file = files.get(ele["key"])
-                        file_content = FileService.parse_docs([upload_file], user_id)
-                        file_name = upload_file.filename
-                        ele["value"] = file_name + "\n" + file_content
-                    else:
-                        if "value" in ele:
-                            ele.pop("value")
-                else:
-                    if req is not None and req.get(ele["key"]):
-                        ele["value"] = req[ele["key"]]
-                    else:
-                        if "value" in ele:
-                            ele.pop("value")
-
-    for ans in canvas.run(stream=False):
-        pass
+    conv = {
+        "id": session_id,
+        "dialog_id": cvs.id,
+        "user_id": user_id,
+        "message": [],
+        "source": "agent",
+        "dsl": cvs.dsl
+    }
+    API4ConversationService.save(**conv)
 
     cvs.dsl = json.loads(str(canvas))
-    conv = {"id": get_uuid(), "dialog_id": cvs.id, "user_id": user_id, "message": [{"role": "assistant", "content": canvas.get_prologue()}], "source": "agent", "dsl": cvs.dsl}
-    API4ConversationService.save(**conv)
+    conv = {"id": session_id, "dialog_id": cvs.id, "user_id": user_id, "message": [{"role": "assistant", "content": canvas.get_prologue()}], "source": "agent", "dsl": cvs.dsl}
     conv["agent_id"] = conv.pop("dialog_id")
     return get_result(data=conv)
 
@@ -589,14 +565,22 @@ def list_agent_session(tenant_id, agent_id):
             if "prompt" in info:
                 info.pop("prompt")
         conv["agent_id"] = conv.pop("dialog_id")
+        # Fix for session listing endpoint
         if conv["reference"]:
             messages = conv["messages"]
             message_num = 0
             chunk_num = 0
+            # Ensure reference is a list type to prevent KeyError
+            if not isinstance(conv["reference"], list):
+                conv["reference"] = []
             while message_num < len(messages):
                 if message_num != 0 and messages[message_num]["role"] != "user":
                     chunk_list = []
-                    if "chunks" in conv["reference"][chunk_num]:
+                    # Add boundary and type checks to prevent KeyError
+                    if (chunk_num < len(conv["reference"]) and
+                            conv["reference"][chunk_num] is not None and
+                            isinstance(conv["reference"][chunk_num], dict) and
+                            "chunks" in conv["reference"][chunk_num]):
                         chunks = conv["reference"][chunk_num]["chunks"]
                         for chunk in chunks:
                             new_chunk = {
@@ -832,6 +816,29 @@ def chatbot_completions(dialog_id):
         return get_result(data=answer)
 
 
+@manager.route("/chatbots/<dialog_id>/info", methods=["GET"])  # noqa: F821
+def chatbots_inputs(dialog_id):
+    token = request.headers.get("Authorization").split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    e, dialog = DialogService.get_by_id(dialog_id)
+    if not e:
+        return get_error_data_result(f"Can't find dialog by ID: {dialog_id}")
+
+    return get_result(
+        data={
+            "title": dialog.name,
+            "avatar": dialog.icon,
+            "prologue": dialog.prompt_config.get("prologue", ""),
+        }
+    )
+
+
 @manager.route("/agentbots/<agent_id>/completions", methods=["POST"])  # noqa: F821
 def agent_bot_completions(agent_id):
     req = request.json
@@ -879,3 +886,279 @@ def begin_inputs(agent_id):
             "prologue": canvas.get_prologue()
         }
     )
+
+
+@manager.route("/searchbots/ask", methods=["POST"])  # noqa: F821
+@validate_request("question", "kb_ids")
+def ask_about_embedded():
+    token = request.headers.get("Authorization").split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    req = request.json
+    uid = objs[0].tenant_id
+
+    search_id = req.get("search_id", "")
+    search_config = {}
+    if search_id:
+        if search_app := SearchService.get_detail(search_id):
+            search_config = search_app.get("search_config", {})
+
+    def stream():
+        nonlocal req, uid
+        try:
+            for ans in ask(req["question"], req["kb_ids"], uid, search_config):
+                yield "data:" + json.dumps({"code": 0, "message": "", "data": ans}, ensure_ascii=False) + "\n\n"
+        except Exception as e:
+            yield "data:" + json.dumps({"code": 500, "message": str(e), "data": {"answer": "**ERROR**: " + str(e), "reference": []}}, ensure_ascii=False) + "\n\n"
+        yield "data:" + json.dumps({"code": 0, "message": "", "data": True}, ensure_ascii=False) + "\n\n"
+
+    resp = Response(stream(), 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
+
+
+@manager.route("/searchbots/retrieval_test", methods=['POST'])  # noqa: F821
+@validate_request("kb_id", "question")
+def retrieval_test_embedded():
+    token = request.headers.get("Authorization").split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    req = request.json
+    page = int(req.get("page", 1))
+    size = int(req.get("size", 30))
+    question = req["question"]
+    kb_ids = req["kb_id"]
+    if isinstance(kb_ids, str):
+        kb_ids = [kb_ids]
+    doc_ids = req.get("doc_ids", [])
+    similarity_threshold = float(req.get("similarity_threshold", 0.0))
+    vector_similarity_weight = float(req.get("vector_similarity_weight", 0.3))
+    use_kg = req.get("use_kg", False)
+    top = int(req.get("top_k", 1024))
+    langs = req.get("cross_languages", [])
+    tenant_ids = []
+
+    tenant_id = objs[0].tenant_id
+    if not tenant_id:
+        return get_error_data_result(message="permission denined.")
+
+    try:
+        tenants = UserTenantService.query(user_id=tenant_id)
+        for kb_id in kb_ids:
+            for tenant in tenants:
+                if KnowledgebaseService.query(
+                        tenant_id=tenant.tenant_id, id=kb_id):
+                    tenant_ids.append(tenant.tenant_id)
+                    break
+            else:
+                return get_json_result(
+                    data=False, message='Only owner of knowledgebase authorized for this operation.',
+                    code=settings.RetCode.OPERATING_ERROR)
+
+        e, kb = KnowledgebaseService.get_by_id(kb_ids[0])
+        if not e:
+            return get_error_data_result(message="Knowledgebase not found!")
+
+        if langs:
+            question = cross_languages(kb.tenant_id, None, question, langs)
+
+        embd_mdl = LLMBundle(kb.tenant_id, LLMType.EMBEDDING.value, llm_name=kb.embd_id)
+
+        rerank_mdl = None
+        if req.get("rerank_id"):
+            rerank_mdl = LLMBundle(kb.tenant_id, LLMType.RERANK.value, llm_name=req["rerank_id"])
+
+        if req.get("keyword", False):
+            chat_mdl = LLMBundle(kb.tenant_id, LLMType.CHAT)
+            question += keyword_extraction(chat_mdl, question)
+
+        labels = label_question(question, [kb])
+        ranks = settings.retrievaler.retrieval(question, embd_mdl, tenant_ids, kb_ids, page, size,
+                               similarity_threshold, vector_similarity_weight, top,
+                               doc_ids, rerank_mdl=rerank_mdl, highlight=req.get("highlight"),
+                               rank_feature=labels
+                               )
+        if use_kg:
+            ck = settings.kg_retrievaler.retrieval(question,
+                                                   tenant_ids,
+                                                   kb_ids,
+                                                   embd_mdl,
+                                                   LLMBundle(kb.tenant_id, LLMType.CHAT))
+            if ck["content_with_weight"]:
+                ranks["chunks"].insert(0, ck)
+
+        for c in ranks["chunks"]:
+            c.pop("vector", None)
+        ranks["labels"] = labels
+
+        return get_json_result(data=ranks)
+    except Exception as e:
+        if str(e).find("not_found") > 0:
+            return get_json_result(data=False, message='No chunk found! Check the chunk status please!',
+                                   code=settings.RetCode.DATA_ERROR)
+        return server_error_response(e)
+
+
+@manager.route("/searchbots/related_questions", methods=["POST"])  # noqa: F821
+@validate_request("question")
+def related_questions_embedded():
+    token = request.headers.get("Authorization").split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    req = request.json
+    tenant_id = objs[0].tenant_id
+    if not tenant_id:
+        return get_error_data_result(message="permission denined.")
+
+    search_id = req.get("search_id", "")
+    search_config = {}
+    if search_id:
+        if search_app := SearchService.get_detail(search_id):
+            search_config = search_app.get("search_config", {})
+
+    question = req["question"]
+
+    chat_id = search_config.get("chat_id", "")
+    chat_mdl = LLMBundle(tenant_id, LLMType.CHAT, chat_id)
+
+    gen_conf = search_config.get("llm_setting", {"temperature": 0.9})
+    prompt = load_prompt("related_question")
+    ans = chat_mdl.chat(
+        prompt,
+        [
+            {
+                "role": "user",
+                "content": f"""
+Keywords: {question}
+Related search terms:
+    """,
+            }
+        ],
+        gen_conf,
+    )
+    return get_json_result(data=[re.sub(r"^[0-9]\. ", "", a) for a in ans.split("\n") if re.match(r"^[0-9]\. ", a)])
+
+
+@manager.route("/searchbots/detail", methods=["GET"])  # noqa: F821
+def detail_share_embedded():
+    token = request.headers.get("Authorization").split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    search_id = request.args["search_id"]
+    tenant_id = objs[0].tenant_id
+    if not tenant_id:
+        return get_error_data_result(message="permission denined.")
+    try:
+        tenants = UserTenantService.query(user_id=tenant_id)
+        for tenant in tenants:
+            if SearchService.query(tenant_id=tenant.tenant_id, id=search_id):
+                break
+        else:
+            return get_json_result(data=False, message="Has no permission for this operation.", code=settings.RetCode.OPERATING_ERROR)
+
+        search = SearchService.get_detail(search_id)
+        if not search:
+            return get_error_data_result(message="Can't find this Search App!")
+        return get_json_result(data=search)
+    except Exception as e:
+        return server_error_response(e)
+
+
+@manager.route("/searchbots/mindmap", methods=["POST"])  # noqa: F821
+@validate_request("question", "kb_ids")
+def mindmap():
+    token = request.headers.get("Authorization").split()
+    if len(token) != 2:
+        return get_error_data_result(message='Authorization is not valid!"')
+    token = token[1]
+    objs = APIToken.query(beta=token)
+    if not objs:
+        return get_error_data_result(message='Authentication error: API key is invalid!"')
+
+    tenant_id = objs[0].tenant_id
+    req = request.json
+
+    search_id = req.get("search_id", "")
+    search_config = {}
+    if search_id:
+        if search_app := SearchService.get_detail(search_id):
+            search_config = search_app.get("search_config", {})
+
+    kb_ids = req["kb_ids"]
+    if search_config.get("kb_ids", []):
+        kb_ids = search_config.get("kb_ids", [])
+    e, kb = KnowledgebaseService.get_by_id(kb_ids[0])
+    if not e:
+        return get_error_data_result(message="Knowledgebase not found!")
+
+    chat_id = ""
+    similarity_threshold = 0.3,
+    vector_similarity_weight = 0.3,
+    top = 1024,
+    doc_ids = []
+    rerank_id = ""
+    rerank_mdl = None
+
+    if search_config:
+        if search_config.get("chat_id", ""):
+            chat_id = search_config.get("chat_id", "")
+        if search_config.get("similarity_threshold", 0.2):
+            similarity_threshold = search_config.get("similarity_threshold", 0.2)
+        if search_config.get("vector_similarity_weight", 0.3):
+            vector_similarity_weight = search_config.get("vector_similarity_weight", 0.3)
+        if search_config.get("top_k", 1024):
+            top = search_config.get("top_k", 1024)
+        if search_config.get("doc_ids", []):
+            doc_ids = search_config.get("doc_ids", [])
+        if search_config.get("rerank_id", ""):
+            rerank_id = search_config.get("rerank_id", "")
+
+    embd_mdl = LLMBundle(tenant_id, LLMType.EMBEDDING, llm_name=kb.embd_id)
+    chat_mdl = LLMBundle(tenant_id, LLMType.CHAT, llm_name=chat_id)
+    if rerank_id:
+        rerank_mdl = LLMBundle(tenant_id, LLMType.RERANK, rerank_id)
+    question = req["question"]
+    ranks = settings.retrievaler.retrieval(
+        question=question,
+        embd_mdl=embd_mdl,
+        tenant_ids=tenant_id,
+        kb_ids=kb_ids,
+        page=1,
+        page_size=12,
+        similarity_threshold=similarity_threshold,
+        vector_similarity_weight=vector_similarity_weight,
+        top=top,
+        doc_ids=doc_ids,
+        aggs=False,
+        rerank_mdl=rerank_mdl,
+        rank_feature=label_question(question, [kb]),
+    )
+    mindmap = MindMapExtractor(chat_mdl)
+    mind_map = trio.run(mindmap, [c["content_with_weight"] for c in ranks["chunks"]])
+    mind_map = mind_map.output
+    if "error" in mind_map:
+        return server_error_response(Exception(mind_map["error"]))
+    return get_json_result(data=mind_map)

api/apps/search_app.py

@@ -22,7 +22,6 @@ from api.constants import DATASET_NAME_LIMIT
 from api.db import StatusEnum
 from api.db.db_models import DB
 from api.db.services import duplicate_name
-from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.search_service import SearchService
 from api.db.services.user_service import TenantService, UserTenantService
 from api.utils import get_uuid
@@ -47,7 +46,7 @@ def create():
         return get_data_error_result(message="Authorizationd identity.")
 
     search_name = search_name.strip()
-    search_name = duplicate_name(KnowledgebaseService.query, name=search_name, tenant_id=current_user.id, status=StatusEnum.VALID.value)
+    search_name = duplicate_name(SearchService.query, name=search_name, tenant_id=current_user.id, status=StatusEnum.VALID.value)
 
     req["id"] = get_uuid()
     req["name"] = search_name

api/apps/tenant_app.py

@@ -18,12 +18,14 @@ from flask import request
 from flask_login import login_required, current_user
 
 from api import settings
+from api.apps import smtp_mail_server
 from api.db import UserTenantRole, StatusEnum
 from api.db.db_models import UserTenant
 from api.db.services.user_service import UserTenantService, UserService
 
 from api.utils import get_uuid, delta_seconds
 from api.utils.api_utils import get_json_result, validate_request, server_error_response, get_data_error_result
+from api.utils.web_utils import send_invite_email
 
 
 @manager.route("/<tenant_id>/user/list", methods=["GET"])  # noqa: F821
@@ -78,6 +80,20 @@ def create(tenant_id):
         role=UserTenantRole.INVITE,
         status=StatusEnum.VALID.value)
 
+    if smtp_mail_server and settings.SMTP_CONF:
+        from threading import Thread
+
+        user_name = ""
+        _, user = UserService.get_by_id(current_user.id)
+        if user:
+            user_name = user.nickname
+
+        Thread(
+            target=send_invite_email,
+            args=(invite_user_email, settings.MAIL_FRONTEND_URL, tenant_id, user_name or current_user.email),
+            daemon=True
+        ).start()
+
     usr = invite_users[0].to_dict()
     usr = {k: v for k, v in usr.items() if k in ["id", "avatar", "email", "nickname"]}
 

api/apps/user_app.py

@@ -28,7 +28,8 @@ from api.apps.auth import get_auth_client
 from api.db import FileType, UserTenantRole
 from api.db.db_models import TenantLLM
 from api.db.services.file_service import FileService
-from api.db.services.llm_service import LLMService, TenantLLMService
+from api.db.services.llm_service import get_init_tenant_llm
+from api.db.services.tenant_llm_service import TenantLLMService
 from api.db.services.user_service import TenantService, UserService, UserTenantService
 from api.utils import (
     current_timestamp,
@@ -619,33 +620,8 @@ def user_register(user_id, user):
         "size": 0,
         "location": "",
     }
-    tenant_llm = []
-    for llm in LLMService.query(fid=settings.LLM_FACTORY):
-        tenant_llm.append(
-            {
-                "tenant_id": user_id,
-                "llm_factory": settings.LLM_FACTORY,
-                "llm_name": llm.llm_name,
-                "model_type": llm.model_type,
-                "api_key": settings.API_KEY,
-                "api_base": settings.LLM_BASE_URL,
-                "max_tokens": llm.max_tokens if llm.max_tokens else 8192,
-            }
-        )
-    if settings.LIGHTEN != 1:
-        for buildin_embedding_model in settings.BUILTIN_EMBEDDING_MODELS:
-            mdlnm, fid = TenantLLMService.split_model_name_and_factory(buildin_embedding_model)
-            tenant_llm.append(
-                {
-                    "tenant_id": user_id,
-                    "llm_factory": fid,
-                    "llm_name": mdlnm,
-                    "model_type": "embedding",
-                    "api_key": "",
-                    "api_base": "",
-                    "max_tokens": 1024 if buildin_embedding_model == "BAAI/bge-large-zh-v1.5@BAAI" else 512,
-                }
-            )
+
+    tenant_llm = get_init_tenant_llm(user_id)
 
     if not UserService.save(**user):
         return

api/db/db_models.py

@@ -742,8 +742,9 @@ class Dialog(DataBaseModel):
     prompt_type = CharField(max_length=16, null=False, default="simple", help_text="simple|advanced", index=True)
     prompt_config = JSONField(
         null=False,
-        default={"system": "", "prologue": "Hi! I'm your assistant, what can I do for you?", "parameters": [], "empty_response": "Sorry! No relevant content was found in the knowledge base!"},
+        default={"system": "", "prologue": "Hi! I'm your assistant. What can I do for you?", "parameters": [], "empty_response": "Sorry! No relevant content was found in the knowledge base!"},
     )
+    meta_data_filter = JSONField(null=True, default={})
 
     similarity_threshold = FloatField(default=0.2)
     vector_similarity_weight = FloatField(default=0.3)
@@ -871,7 +872,7 @@ class Search(DataBaseModel):
         default={
             "kb_ids": [],
             "doc_ids": [],
-            "similarity_threshold": 0.0,
+            "similarity_threshold": 0.2,
             "vector_similarity_weight": 0.3,
             "use_kg": False,
             # rerank settings
@@ -880,11 +881,12 @@ class Search(DataBaseModel):
             # chat settings
             "summary": False,
             "chat_id": "",
+            # Leave it here for reference, don't need to set default values
             "llm_setting": {
-                "temperature": 0.1,
-                "top_p": 0.3,
-                "frequency_penalty": 0.7,
-                "presence_penalty": 0.4,
+                # "temperature": 0.1,
+                # "top_p": 0.3,
+                # "frequency_penalty": 0.7,
+                # "presence_penalty": 0.4,
             },
             "chat_settingcross_languages": [],
             "highlight": False,
@@ -1015,4 +1017,8 @@ def migrate_db():
         migrate(migrator.add_column("api_4_conversation", "errors", TextField(null=True, help_text="errors")))
     except Exception:
         pass
-    logging.disable(logging.NOTSET)
+    try:
+        migrate(migrator.add_column("dialog", "meta_data_filter", JSONField(null=True, default={})))
+    except Exception:
+        pass
+    logging.disable(logging.NOTSET)

api/db/init_data.py

@@ -27,7 +27,8 @@ from api.db.services import UserService
 from api.db.services.canvas_service import CanvasTemplateService
 from api.db.services.document_service import DocumentService
 from api.db.services.knowledgebase_service import KnowledgebaseService
-from api.db.services.llm_service import LLMFactoriesService, LLMService, TenantLLMService, LLMBundle
+from api.db.services.tenant_llm_service import LLMFactoriesService, TenantLLMService
+from api.db.services.llm_service import LLMService, LLMBundle, get_init_tenant_llm
 from api.db.services.user_service import TenantService, UserTenantService
 from api import settings
 from api.utils.file_utils import get_project_base_directory
@@ -63,12 +64,8 @@ def init_superuser():
         "invited_by": user_info["id"],
         "role": UserTenantRole.OWNER
     }
-    tenant_llm = []
-    for llm in LLMService.query(fid=settings.LLM_FACTORY):
-        tenant_llm.append(
-            {"tenant_id": user_info["id"], "llm_factory": settings.LLM_FACTORY, "llm_name": llm.llm_name,
-             "model_type": llm.model_type,
-             "api_key": settings.API_KEY, "api_base": settings.LLM_BASE_URL})
+
+    tenant_llm = get_init_tenant_llm(user_info["id"])
 
     if not UserService.save(**user_info):
         logging.error("can't init admin.")
@@ -103,7 +100,7 @@ def init_llm_factory():
     except Exception:
         pass
 
-    factory_llm_infos = settings.FACTORY_LLM_INFOS    
+    factory_llm_infos = settings.FACTORY_LLM_INFOS
     for factory_llm_info in factory_llm_infos:
         info = deepcopy(factory_llm_info)
         llm_infos = info.pop("llm")

api/db/services/dialog_service.py

@@ -30,14 +30,17 @@ from api import settings
 from api.db import LLMType, ParserType, StatusEnum
 from api.db.db_models import DB, Dialog
 from api.db.services.common_service import CommonService
+from api.db.services.document_service import DocumentService
 from api.db.services.knowledgebase_service import KnowledgebaseService
 from api.db.services.langfuse_service import TenantLangfuseService
-from api.db.services.llm_service import LLMBundle, TenantLLMService
+from api.db.services.llm_service import LLMBundle
+from api.db.services.tenant_llm_service import TenantLLMService
 from api.utils import current_timestamp, datetime_format
 from rag.app.resume import forbidden_select_fields4resume
 from rag.app.tag import label_question
 from rag.nlp.search import index_name
 from rag.prompts import chunks_format, citation_prompt, cross_languages, full_question, kb_prompt, keyword_extraction, message_fit_in
+from rag.prompts.prompts import gen_meta_filter, PROMPT_JINJA_ENV, ASK_SUMMARY
 from rag.utils import num_tokens_from_string, rmSpace
 from rag.utils.tavily_conn import Tavily
 
@@ -96,7 +99,6 @@ 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):
@@ -119,6 +121,7 @@ class DialogService(CommonService):
             cls.model.do_refer,
             cls.model.rerank_id,
             cls.model.kb_ids,
+            cls.model.icon,
             cls.model.status,
             User.nickname,
             User.avatar.alias("tenant_avatar"),
@@ -250,6 +253,47 @@ def repair_bad_citation_formats(answer: str, kbinfos: dict, idx: set):
     return answer, idx
 
 
+def meta_filter(metas: dict, filters: list[dict]):
+    doc_ids = []
+
+    def filter_out(v2docs, operator, value):
+        nonlocal doc_ids
+        for input, docids in v2docs.items():
+            try:
+                input = float(input)
+                value = float(value)
+            except Exception:
+                input = str(input)
+                value = str(value)
+
+            for conds in [
+                    (operator == "contains", str(value).lower() in str(input).lower()),
+                    (operator == "not contains", str(value).lower() not in str(input).lower()),
+                    (operator == "start with", str(input).lower().startswith(str(value).lower())),
+                    (operator == "end with", str(input).lower().endswith(str(value).lower())),
+                    (operator == "empty", not input),
+                    (operator == "not empty", input),
+                    (operator == "=", input == value),
+                    (operator == "≠", input != value),
+                    (operator == ">", input > value),
+                    (operator == "<", input < value),
+                    (operator == "≥", input >= value),
+                    (operator == "≤", input <= value),
+                ]:
+                try:
+                    if all(conds):
+                        doc_ids.extend(docids)
+                except Exception:
+                    pass
+
+    for k, v2docs in metas.items():
+        for f in filters:
+            if k != f["key"]:
+                continue
+            filter_out(v2docs, f["op"], f["value"])
+    return doc_ids
+
+
 def chat(dialog, messages, stream=True, **kwargs):
     assert messages[-1]["role"] == "user", "The last content of this conversation is not from user."
     if not dialog.kb_ids and not dialog.prompt_config.get("tavily_api_key"):
@@ -287,9 +331,10 @@ def chat(dialog, messages, stream=True, **kwargs):
 
     retriever = settings.retrievaler
     questions = [m["content"] for m in messages if m["role"] == "user"][-3:]
-    attachments = kwargs["doc_ids"].split(",") if "doc_ids" in kwargs else None
+    attachments = kwargs["doc_ids"].split(",") if "doc_ids" in kwargs else []
     if "doc_ids" in messages[-1]:
         attachments = messages[-1]["doc_ids"]
+
     prompt_config = dialog.prompt_config
     field_map = KnowledgebaseService.get_field_map(dialog.kb_ids)
     # try to use sql if field mapping is good to go
@@ -316,6 +361,18 @@ def chat(dialog, messages, stream=True, **kwargs):
     if prompt_config.get("cross_languages"):
         questions = [cross_languages(dialog.tenant_id, dialog.llm_id, questions[0], prompt_config["cross_languages"])]
 
+    if dialog.meta_data_filter:
+        metas = DocumentService.get_meta_by_kbs(dialog.kb_ids)
+        if dialog.meta_data_filter.get("method") == "auto":
+            filters = gen_meta_filter(chat_mdl, metas, questions[-1])
+            attachments.extend(meta_filter(metas, filters))
+            if not attachments:
+                attachments = None
+        elif dialog.meta_data_filter.get("method") == "manual":
+            attachments.extend(meta_filter(metas, dialog.meta_data_filter["manual"]))
+            if not attachments:
+                attachments = None
+
     if prompt_config.get("keyword", False):
         questions[-1] += keyword_extraction(chat_mdl, questions[-1])
 
@@ -323,17 +380,26 @@ def chat(dialog, messages, stream=True, **kwargs):
 
     thought = ""
     kbinfos = {"total": 0, "chunks": [], "doc_aggs": []}
+    knowledges = []
 
-    if "knowledge" not in [p["key"] for p in prompt_config["parameters"]]:
-        knowledges = []
-    else:
+    if attachments is not None and "knowledge" in [p["key"] for p in prompt_config["parameters"]]:
         tenant_ids = list(set([kb.tenant_id for kb in kbs]))
         knowledges = []
         if prompt_config.get("reasoning", False):
             reasoner = DeepResearcher(
                 chat_mdl,
                 prompt_config,
-                partial(retriever.retrieval, embd_mdl=embd_mdl, tenant_ids=tenant_ids, kb_ids=dialog.kb_ids, page=1, page_size=dialog.top_n, similarity_threshold=0.2, vector_similarity_weight=0.3),
+                partial(
+                    retriever.retrieval,
+                    embd_mdl=embd_mdl,
+                    tenant_ids=tenant_ids,
+                    kb_ids=dialog.kb_ids,
+                    page=1,
+                    page_size=dialog.top_n,
+                    similarity_threshold=0.2,
+                    vector_similarity_weight=0.3,
+                    doc_ids=attachments,
+                ),
             )
 
             for think in reasoner.thinking(kbinfos, " ".join(questions)):
@@ -621,7 +687,30 @@ def tts(tts_mdl, text):
     return binascii.hexlify(bin).decode("utf-8")
 
 
-def ask(question, kb_ids, tenant_id, chat_llm_name=None):
+def ask(question, kb_ids, tenant_id, chat_llm_name=None, search_config={}):
+    similarity_threshold = 0.1,
+    vector_similarity_weight = 0.3,
+    top = 1024,
+    doc_ids = []
+    rerank_id = ""
+    rerank_mdl = None
+
+    if search_config:
+        if search_config.get("kb_ids", []):
+            kb_ids = search_config.get("kb_ids", [])
+        if search_config.get("chat_id", ""):
+            chat_llm_name = search_config.get("chat_id", "")
+        if search_config.get("similarity_threshold", 0.1):
+            similarity_threshold = search_config.get("similarity_threshold", 0.1)
+        if search_config.get("vector_similarity_weight", 0.3):
+            vector_similarity_weight = search_config.get("vector_similarity_weight", 0.3)
+        if search_config.get("top_k", 1024):
+            top = search_config.get("top_k", 1024)
+        if search_config.get("doc_ids", []):
+            doc_ids = search_config.get("doc_ids", [])
+        if search_config.get("rerank_id", ""):
+            rerank_id = search_config.get("rerank_id", "")
+
     kbs = KnowledgebaseService.get_by_ids(kb_ids)
     embedding_list = list(set([kb.embd_id for kb in kbs]))
 
@@ -630,30 +719,34 @@ def ask(question, kb_ids, tenant_id, chat_llm_name=None):
 
     embd_mdl = LLMBundle(tenant_id, LLMType.EMBEDDING, embedding_list[0])
     chat_mdl = LLMBundle(tenant_id, LLMType.CHAT, chat_llm_name)
+    if rerank_id:
+        rerank_mdl = LLMBundle(tenant_id, LLMType.RERANK, rerank_id)
     max_tokens = chat_mdl.max_length
     tenant_ids = list(set([kb.tenant_id for kb in kbs]))
-    kbinfos = retriever.retrieval(question, embd_mdl, tenant_ids, kb_ids, 1, 12, 0.1, 0.3, aggs=False, rank_feature=label_question(question, kbs))
+
+    kbinfos = retriever.retrieval(
+        question = question,
+        embd_mdl=embd_mdl,
+        tenant_ids=tenant_ids,
+        kb_ids=kb_ids,
+        page=1,
+        page_size=12,
+        similarity_threshold=similarity_threshold,
+        vector_similarity_weight=vector_similarity_weight,
+        top=top,
+        doc_ids=doc_ids,
+        aggs=False,
+        rerank_mdl=rerank_mdl,
+        rank_feature=label_question(question, kbs)
+    )
+
     knowledges = kb_prompt(kbinfos, max_tokens)
-    prompt = """
-    Role: You're a smart assistant. Your name is Miss R.
-    Task: Summarize the information from knowledge bases and answer user's question.
-    Requirements and restriction:
-      - DO NOT make things up, especially for numbers.
-      - If the information from knowledge is irrelevant with user's question, JUST SAY: Sorry, no relevant information provided.
-      - Answer with markdown format text.
-      - Answer in language of user's question.
-      - DO NOT make things up, especially for numbers.
+    sys_prompt = PROMPT_JINJA_ENV.from_string(ASK_SUMMARY).render(knowledge="\n".join(knowledges))
 
-    ### Information from knowledge bases
-    %s
-
-    The above is information from knowledge bases.
-
-    """ % "\n".join(knowledges)
     msg = [{"role": "user", "content": question}]
 
     def decorate_answer(answer):
-        nonlocal knowledges, kbinfos, prompt
+        nonlocal knowledges, kbinfos, sys_prompt
         answer, idx = retriever.insert_citations(answer, [ck["content_ltks"] for ck in kbinfos["chunks"]], [ck["vector"] for ck in kbinfos["chunks"]], embd_mdl, tkweight=0.7, vtweight=0.3)
         idx = set([kbinfos["chunks"][int(i)]["doc_id"] for i in idx])
         recall_docs = [d for d in kbinfos["doc_aggs"] if d["doc_id"] in idx]
@@ -671,7 +764,7 @@ def ask(question, kb_ids, tenant_id, chat_llm_name=None):
         return {"answer": answer, "reference": refs}
 
     answer = ""
-    for ans in chat_mdl.chat_streamly(prompt, msg, {"temperature": 0.1}):
+    for ans in chat_mdl.chat_streamly(sys_prompt, msg, {"temperature": 0.1}):
         answer = ans
         yield {"answer": answer, "reference": {}}
     yield decorate_answer(answer)

api/db/services/document_service.py

@@ -243,7 +243,7 @@ class DocumentService(CommonService):
         from api.db.services.task_service import TaskService
         cls.clear_chunk_num(doc.id)
         try:
-            TaskService.filter_delete(Task.doc_id == doc.id)
+            TaskService.filter_delete([Task.doc_id == doc.id])
             page = 0
             page_size = 1000
             all_chunk_ids = []
@@ -574,6 +574,25 @@ class DocumentService(CommonService):
     def update_meta_fields(cls, doc_id, meta_fields):
         return cls.update_by_id(doc_id, {"meta_fields": meta_fields})
 
+    @classmethod
+    @DB.connection_context()
+    def get_meta_by_kbs(cls, kb_ids):
+        fields = [
+            cls.model.id,
+            cls.model.meta_fields,
+        ]
+        meta = {}
+        for r in cls.model.select(*fields).where(cls.model.kb_id.in_(kb_ids)):
+            doc_id = r.id
+            for k,v in r.meta_fields.items():
+                if k not in meta:
+                    meta[k] = {}
+                v = str(v)
+                if v not in meta[k]:
+                    meta[k][v] = []
+                meta[k][v].append(doc_id)
+        return meta
+
     @classmethod
     @DB.connection_context()
     def update_progress(cls):

api/db/services/file_service.py

@@ -227,10 +227,13 @@ class FileService(CommonService):
         #     tenant_id: Tenant ID
         # Returns:
         #     Knowledge base folder dictionary
-        for root in cls.model.select().where((cls.model.tenant_id == tenant_id), (cls.model.parent_id == cls.model.id)):
-            for folder in cls.model.select().where((cls.model.tenant_id == tenant_id), (cls.model.parent_id == root.id), (cls.model.name == KNOWLEDGEBASE_FOLDER_NAME)):
-                return folder.to_dict()
-        assert False, "Can't find the KB folder. Database init error."
+        root_folder = cls.get_root_folder(tenant_id)
+        root_id = root_folder["id"]
+        kb_folder = cls.model.select().where((cls.model.tenant_id == tenant_id), (cls.model.parent_id == root_id), (cls.model.name == KNOWLEDGEBASE_FOLDER_NAME)).first()
+        if not kb_folder:
+            kb_folder = cls.new_a_file_from_kb(tenant_id, KNOWLEDGEBASE_FOLDER_NAME, root_id)
+            return kb_folder
+        return kb_folder.to_dict()
 
     @classmethod
     @DB.connection_context()
@@ -499,10 +502,9 @@ class FileService(CommonService):
     @staticmethod
     def get_blob(user_id, location):
         bname = f"{user_id}-downloads"
-        return  STORAGE_IMPL.get(bname, location)
+        return STORAGE_IMPL.get(bname, location)
 
     @staticmethod
     def put_blob(user_id, location, blob):
         bname = f"{user_id}-downloads"
-        return  STORAGE_IMPL.put(bname, location, blob)
-
+        return STORAGE_IMPL.put(bname, location, blob)

api/db/services/llm_service.py

@@ -13,249 +13,78 @@
 #  See the License for the specific language governing permissions and
 #  limitations under the License.
 #
+import inspect
 import logging
 import re
 from functools import partial
 from typing import Generator
-
-from langfuse import Langfuse
-
-from api import settings
-from api.db import LLMType
-from api.db.db_models import DB, LLM, LLMFactories, TenantLLM
+from api.db.db_models import LLM
 from api.db.services.common_service import CommonService
-from api.db.services.langfuse_service import TenantLangfuseService
-from api.db.services.user_service import TenantService
-from rag.llm import ChatModel, CvModel, EmbeddingModel, RerankModel, Seq2txtModel, TTSModel
-
-
-class LLMFactoriesService(CommonService):
-    model = LLMFactories
+from api.db.services.tenant_llm_service import LLM4Tenant, TenantLLMService
 
 
 class LLMService(CommonService):
     model = LLM
 
 
-class TenantLLMService(CommonService):
-    model = TenantLLM
+def get_init_tenant_llm(user_id):
+    from api import settings
+    tenant_llm = []
 
-    @classmethod
-    @DB.connection_context()
-    def get_api_key(cls, tenant_id, model_name):
-        mdlnm, fid = TenantLLMService.split_model_name_and_factory(model_name)
-        if not fid:
-            objs = cls.query(tenant_id=tenant_id, llm_name=mdlnm)
-        else:
-            objs = cls.query(tenant_id=tenant_id, llm_name=mdlnm, llm_factory=fid)
+    seen = set()
+    factory_configs = []
+    for factory_config in [
+        settings.CHAT_CFG,
+        settings.EMBEDDING_CFG,
+        settings.ASR_CFG,
+        settings.IMAGE2TEXT_CFG,
+        settings.RERANK_CFG,
+    ]:
+        factory_name = factory_config["factory"]
+        if factory_name not in seen:
+            seen.add(factory_name)
+            factory_configs.append(factory_config)
 
-        if (not objs) and fid:
-            if fid == "LocalAI":
-                mdlnm += "___LocalAI"
-            elif fid == "HuggingFace":
-                mdlnm += "___HuggingFace"
-            elif fid == "OpenAI-API-Compatible":
-                mdlnm += "___OpenAI-API"
-            elif fid == "VLLM":
-                mdlnm += "___VLLM"
-            objs = cls.query(tenant_id=tenant_id, llm_name=mdlnm, llm_factory=fid)
-        if not objs:
-            return
-        return objs[0]
-
-    @classmethod
-    @DB.connection_context()
-    def get_my_llms(cls, tenant_id):
-        fields = [cls.model.llm_factory, LLMFactories.logo, LLMFactories.tags, cls.model.model_type, cls.model.llm_name, cls.model.used_tokens]
-        objs = cls.model.select(*fields).join(LLMFactories, on=(cls.model.llm_factory == LLMFactories.name)).where(cls.model.tenant_id == tenant_id, ~cls.model.api_key.is_null()).dicts()
-
-        return list(objs)
-
-    @staticmethod
-    def split_model_name_and_factory(model_name):
-        arr = model_name.split("@")
-        if len(arr) < 2:
-            return model_name, None
-        if len(arr) > 2:
-            return "@".join(arr[0:-1]), arr[-1]
-
-        # model name must be xxx@yyy
-        try:
-            model_factories = settings.FACTORY_LLM_INFOS
-            model_providers = set([f["name"] for f in model_factories])
-            if arr[-1] not in model_providers:
-                return model_name, None
-            return arr[0], arr[-1]
-        except Exception as e:
-            logging.exception(f"TenantLLMService.split_model_name_and_factory got exception: {e}")
-        return model_name, None
-
-    @classmethod
-    @DB.connection_context()
-    def get_model_config(cls, tenant_id, llm_type, llm_name=None):
-        e, tenant = TenantService.get_by_id(tenant_id)
-        if not e:
-            raise LookupError("Tenant not found")
-
-        if llm_type == LLMType.EMBEDDING.value:
-            mdlnm = tenant.embd_id if not llm_name else llm_name
-        elif llm_type == LLMType.SPEECH2TEXT.value:
-            mdlnm = tenant.asr_id
-        elif llm_type == LLMType.IMAGE2TEXT.value:
-            mdlnm = tenant.img2txt_id if not llm_name else llm_name
-        elif llm_type == LLMType.CHAT.value:
-            mdlnm = tenant.llm_id if not llm_name else llm_name
-        elif llm_type == LLMType.RERANK:
-            mdlnm = tenant.rerank_id if not llm_name else llm_name
-        elif llm_type == LLMType.TTS:
-            mdlnm = tenant.tts_id if not llm_name else llm_name
-        else:
-            assert False, "LLM type error"
-
-        model_config = cls.get_api_key(tenant_id, mdlnm)
-        mdlnm, fid = TenantLLMService.split_model_name_and_factory(mdlnm)
-        if not model_config:  # for some cases seems fid mismatch
-            model_config = cls.get_api_key(tenant_id, mdlnm)
-        if model_config:
-            model_config = model_config.to_dict()
-            llm = LLMService.query(llm_name=mdlnm) if not fid else LLMService.query(llm_name=mdlnm, fid=fid)
-            if not llm and fid:  # for some cases seems fid mismatch
-                llm = LLMService.query(llm_name=mdlnm)
-            if llm:
-                model_config["is_tools"] = llm[0].is_tools
-        if not model_config:
-            if llm_type in [LLMType.EMBEDDING, LLMType.RERANK]:
-                llm = LLMService.query(llm_name=mdlnm) if not fid else LLMService.query(llm_name=mdlnm, fid=fid)
-                if llm and llm[0].fid in ["Youdao", "FastEmbed", "BAAI"]:
-                    model_config = {"llm_factory": llm[0].fid, "api_key": "", "llm_name": mdlnm, "api_base": ""}
-            if not model_config:
-                if mdlnm == "flag-embedding":
-                    model_config = {"llm_factory": "Tongyi-Qianwen", "api_key": "", "llm_name": llm_name, "api_base": ""}
-                else:
-                    if not mdlnm:
-                        raise LookupError(f"Type of {llm_type} model is not set.")
-                    raise LookupError("Model({}) not authorized".format(mdlnm))
-        return model_config
-
-    @classmethod
-    @DB.connection_context()
-    def model_instance(cls, tenant_id, llm_type, llm_name=None, lang="Chinese", **kwargs):
-        model_config = TenantLLMService.get_model_config(tenant_id, llm_type, llm_name)
-        if llm_type == LLMType.EMBEDDING.value:
-            if model_config["llm_factory"] not in EmbeddingModel:
-                return
-            return EmbeddingModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], base_url=model_config["api_base"])
-
-        if llm_type == LLMType.RERANK:
-            if model_config["llm_factory"] not in RerankModel:
-                return
-            return RerankModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], base_url=model_config["api_base"])
-
-        if llm_type == LLMType.IMAGE2TEXT.value:
-            if model_config["llm_factory"] not in CvModel:
-                return
-            return CvModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], lang, base_url=model_config["api_base"], **kwargs)
-
-        if llm_type == LLMType.CHAT.value:
-            if model_config["llm_factory"] not in ChatModel:
-                return
-            return ChatModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], base_url=model_config["api_base"], **kwargs)
-
-        if llm_type == LLMType.SPEECH2TEXT:
-            if model_config["llm_factory"] not in Seq2txtModel:
-                return
-            return Seq2txtModel[model_config["llm_factory"]](key=model_config["api_key"], model_name=model_config["llm_name"], lang=lang, base_url=model_config["api_base"])
-        if llm_type == LLMType.TTS:
-            if model_config["llm_factory"] not in TTSModel:
-                return
-            return TTSModel[model_config["llm_factory"]](
-                model_config["api_key"],
-                model_config["llm_name"],
-                base_url=model_config["api_base"],
+    for factory_config in factory_configs:
+        for llm in LLMService.query(fid=factory_config["factory"]):
+            tenant_llm.append(
+                {
+                    "tenant_id": user_id,
+                    "llm_factory": factory_config["factory"],
+                    "llm_name": llm.llm_name,
+                    "model_type": llm.model_type,
+                    "api_key": factory_config["api_key"],
+                    "api_base": factory_config["base_url"],
+                    "max_tokens": llm.max_tokens if llm.max_tokens else 8192,
+                }
             )
 
-    @classmethod
-    @DB.connection_context()
-    def increase_usage(cls, tenant_id, llm_type, used_tokens, llm_name=None):
-        e, tenant = TenantService.get_by_id(tenant_id)
-        if not e:
-            logging.error(f"Tenant not found: {tenant_id}")
-            return 0
-
-        llm_map = {
-            LLMType.EMBEDDING.value: tenant.embd_id if not llm_name else llm_name,
-            LLMType.SPEECH2TEXT.value: tenant.asr_id,
-            LLMType.IMAGE2TEXT.value: tenant.img2txt_id,
-            LLMType.CHAT.value: tenant.llm_id if not llm_name else llm_name,
-            LLMType.RERANK.value: tenant.rerank_id if not llm_name else llm_name,
-            LLMType.TTS.value: tenant.tts_id if not llm_name else llm_name,
-        }
-
-        mdlnm = llm_map.get(llm_type)
-        if mdlnm is None:
-            logging.error(f"LLM type error: {llm_type}")
-            return 0
-
-        llm_name, llm_factory = TenantLLMService.split_model_name_and_factory(mdlnm)
-
-        try:
-            num = (
-                cls.model.update(used_tokens=cls.model.used_tokens + used_tokens)
-                .where(cls.model.tenant_id == tenant_id, cls.model.llm_name == llm_name, cls.model.llm_factory == llm_factory if llm_factory else True)
-                .execute()
+    if settings.LIGHTEN != 1:
+        for buildin_embedding_model in settings.BUILTIN_EMBEDDING_MODELS:
+            mdlnm, fid = TenantLLMService.split_model_name_and_factory(buildin_embedding_model)
+            tenant_llm.append(
+                {
+                    "tenant_id": user_id,
+                    "llm_factory": fid,
+                    "llm_name": mdlnm,
+                    "model_type": "embedding",
+                    "api_key": "",
+                    "api_base": "",
+                    "max_tokens": 1024 if buildin_embedding_model == "BAAI/bge-large-zh-v1.5@BAAI" else 512,
+                }
             )
-        except Exception:
-            logging.exception("TenantLLMService.increase_usage got exception,Failed to update used_tokens for tenant_id=%s, llm_name=%s", tenant_id, llm_name)
-            return 0
 
-        return num
-
-    @classmethod
-    @DB.connection_context()
-    def get_openai_models(cls):
-        objs = cls.model.select().where((cls.model.llm_factory == "OpenAI"), ~(cls.model.llm_name == "text-embedding-3-small"), ~(cls.model.llm_name == "text-embedding-3-large")).dicts()
-        return list(objs)
-
-    @staticmethod
-    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:
-            for llm in llm_factory["llm"]:
-                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
-
-        llm = TenantLLMService.get_or_none(llm_name=llm_id)
-        if llm:
-            return llm.model_type
-        for llm in TenantLLMService.query(llm_name=llm_id):
-            return llm.model_type
+    unique = {}
+    for item in tenant_llm:
+        key = (item["tenant_id"], item["llm_factory"], item["llm_name"])
+        if key not in unique:
+            unique[key] = item
+    return list(unique.values())
 
 
-class LLMBundle:
+class LLMBundle(LLM4Tenant):
     def __init__(self, tenant_id, llm_type, llm_name=None, lang="Chinese", **kwargs):
-        self.tenant_id = tenant_id
-        self.llm_type = llm_type
-        self.llm_name = llm_name
-        self.mdl = TenantLLMService.model_instance(tenant_id, llm_type, llm_name, lang=lang, **kwargs)
-        assert self.mdl, "Can't find model for {}/{}/{}".format(tenant_id, llm_type, llm_name)
-        model_config = TenantLLMService.get_model_config(tenant_id, llm_type, llm_name)
-        self.max_length = model_config.get("max_tokens", 8192)
-
-        self.is_tools = model_config.get("is_tools", False)
-        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
-                trace_id = self.langfuse.create_trace_id()
-                self.trace_context = {"trace_id": trace_id}
+        super().__init__(tenant_id, llm_type, llm_name, lang, **kwargs)
 
     def bind_tools(self, toolcall_session, tools):
         if not self.is_tools:
@@ -376,7 +205,24 @@ class LLMBundle:
             return txt
 
         return txt[last_think_end + len("</think>") :]
+    
+    @staticmethod
+    def _clean_param(chat_partial, **kwargs):
+        func = chat_partial.func
+        sig = inspect.signature(func)
+        keyword_args = []
+        support_var_args = False
+        for param in sig.parameters.values():
+            if param.kind == inspect.Parameter.VAR_KEYWORD or param.kind == inspect.Parameter.VAR_POSITIONAL:
+                support_var_args = True
+            elif param.kind == inspect.Parameter.KEYWORD_ONLY:
+                keyword_args.append(param.name)
 
+        use_kwargs = kwargs
+        if not support_var_args:
+            use_kwargs = {k: v for k, v in kwargs.items() if k in keyword_args}
+        return use_kwargs
+        
     def chat(self, system: str, history: list, gen_conf: dict = {}, **kwargs) -> str:
         if self.langfuse:
             generation = self.langfuse.start_generation(trace_context=self.trace_context, name="chat", model=self.llm_name, input={"system": system, "history": history})
@@ -384,8 +230,9 @@ class LLMBundle:
         chat_partial = partial(self.mdl.chat, system, history, gen_conf)
         if self.is_tools and self.mdl.is_tools:
             chat_partial = partial(self.mdl.chat_with_tools, system, history, gen_conf)
-
-        txt, used_tokens = chat_partial(**kwargs)
+            
+        use_kwargs = self._clean_param(chat_partial, **kwargs)
+        txt, used_tokens = chat_partial(**use_kwargs)
         txt = self._remove_reasoning_content(txt)
 
         if not self.verbose_tool_use:
@@ -409,8 +256,8 @@ class LLMBundle:
         total_tokens = 0
         if self.is_tools and self.mdl.is_tools:
             chat_partial = partial(self.mdl.chat_streamly_with_tools, system, history, gen_conf)
-
-        for txt in chat_partial(**kwargs):
+        use_kwargs = self._clean_param(chat_partial, **kwargs)
+        for txt in chat_partial(**use_kwargs):
             if isinstance(txt, int):
                 total_tokens = txt
                 if self.langfuse:

api/db/services/tenant_llm_service.py

@@ -0,0 +1,252 @@
+#
+#  Copyright 2024 The InfiniFlow Authors. All Rights Reserved.
+#
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+#
+import logging
+from langfuse import Langfuse
+from api import settings
+from api.db import LLMType
+from api.db.db_models import DB, LLMFactories, TenantLLM
+from api.db.services.common_service import CommonService
+from api.db.services.langfuse_service import TenantLangfuseService
+from api.db.services.user_service import TenantService
+from rag.llm import ChatModel, CvModel, EmbeddingModel, RerankModel, Seq2txtModel, TTSModel
+
+
+class LLMFactoriesService(CommonService):
+    model = LLMFactories
+
+
+class TenantLLMService(CommonService):
+    model = TenantLLM
+
+    @classmethod
+    @DB.connection_context()
+    def get_api_key(cls, tenant_id, model_name):
+        mdlnm, fid = TenantLLMService.split_model_name_and_factory(model_name)
+        if not fid:
+            objs = cls.query(tenant_id=tenant_id, llm_name=mdlnm)
+        else:
+            objs = cls.query(tenant_id=tenant_id, llm_name=mdlnm, llm_factory=fid)
+
+        if (not objs) and fid:
+            if fid == "LocalAI":
+                mdlnm += "___LocalAI"
+            elif fid == "HuggingFace":
+                mdlnm += "___HuggingFace"
+            elif fid == "OpenAI-API-Compatible":
+                mdlnm += "___OpenAI-API"
+            elif fid == "VLLM":
+                mdlnm += "___VLLM"
+            objs = cls.query(tenant_id=tenant_id, llm_name=mdlnm, llm_factory=fid)
+        if not objs:
+            return
+        return objs[0]
+
+    @classmethod
+    @DB.connection_context()
+    def get_my_llms(cls, tenant_id):
+        fields = [cls.model.llm_factory, LLMFactories.logo, LLMFactories.tags, cls.model.model_type, cls.model.llm_name, cls.model.used_tokens]
+        objs = cls.model.select(*fields).join(LLMFactories, on=(cls.model.llm_factory == LLMFactories.name)).where(cls.model.tenant_id == tenant_id, ~cls.model.api_key.is_null()).dicts()
+
+        return list(objs)
+
+    @staticmethod
+    def split_model_name_and_factory(model_name):
+        arr = model_name.split("@")
+        if len(arr) < 2:
+            return model_name, None
+        if len(arr) > 2:
+            return "@".join(arr[0:-1]), arr[-1]
+
+        # model name must be xxx@yyy
+        try:
+            model_factories = settings.FACTORY_LLM_INFOS
+            model_providers = set([f["name"] for f in model_factories])
+            if arr[-1] not in model_providers:
+                return model_name, None
+            return arr[0], arr[-1]
+        except Exception as e:
+            logging.exception(f"TenantLLMService.split_model_name_and_factory got exception: {e}")
+        return model_name, None
+
+    @classmethod
+    @DB.connection_context()
+    def get_model_config(cls, tenant_id, llm_type, llm_name=None):
+        from api.db.services.llm_service import LLMService
+        e, tenant = TenantService.get_by_id(tenant_id)
+        if not e:
+            raise LookupError("Tenant not found")
+
+        if llm_type == LLMType.EMBEDDING.value:
+            mdlnm = tenant.embd_id if not llm_name else llm_name
+        elif llm_type == LLMType.SPEECH2TEXT.value:
+            mdlnm = tenant.asr_id
+        elif llm_type == LLMType.IMAGE2TEXT.value:
+            mdlnm = tenant.img2txt_id if not llm_name else llm_name
+        elif llm_type == LLMType.CHAT.value:
+            mdlnm = tenant.llm_id if not llm_name else llm_name
+        elif llm_type == LLMType.RERANK:
+            mdlnm = tenant.rerank_id if not llm_name else llm_name
+        elif llm_type == LLMType.TTS:
+            mdlnm = tenant.tts_id if not llm_name else llm_name
+        else:
+            assert False, "LLM type error"
+
+        model_config = cls.get_api_key(tenant_id, mdlnm)
+        mdlnm, fid = TenantLLMService.split_model_name_and_factory(mdlnm)
+        if not model_config:  # for some cases seems fid mismatch
+            model_config = cls.get_api_key(tenant_id, mdlnm)
+        if model_config:
+            model_config = model_config.to_dict()
+            llm = LLMService.query(llm_name=mdlnm) if not fid else LLMService.query(llm_name=mdlnm, fid=fid)
+            if not llm and fid:  # for some cases seems fid mismatch
+                llm = LLMService.query(llm_name=mdlnm)
+            if llm:
+                model_config["is_tools"] = llm[0].is_tools
+        if not model_config:
+            if llm_type in [LLMType.EMBEDDING, LLMType.RERANK]:
+                llm = LLMService.query(llm_name=mdlnm) if not fid else LLMService.query(llm_name=mdlnm, fid=fid)
+                if llm and llm[0].fid in ["Youdao", "FastEmbed", "BAAI"]:
+                    model_config = {"llm_factory": llm[0].fid, "api_key": "", "llm_name": mdlnm, "api_base": ""}
+            if not model_config:
+                if mdlnm == "flag-embedding":
+                    model_config = {"llm_factory": "Tongyi-Qianwen", "api_key": "", "llm_name": llm_name, "api_base": ""}
+                else:
+                    if not mdlnm:
+                        raise LookupError(f"Type of {llm_type} model is not set.")
+                    raise LookupError("Model({}) not authorized".format(mdlnm))
+        return model_config
+
+    @classmethod
+    @DB.connection_context()
+    def model_instance(cls, tenant_id, llm_type, llm_name=None, lang="Chinese", **kwargs):
+        model_config = TenantLLMService.get_model_config(tenant_id, llm_type, llm_name)
+        kwargs.update({"provider": model_config["llm_factory"]})
+        if llm_type == LLMType.EMBEDDING.value:
+            if model_config["llm_factory"] not in EmbeddingModel:
+                return
+            return EmbeddingModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], base_url=model_config["api_base"])
+
+        if llm_type == LLMType.RERANK:
+            if model_config["llm_factory"] not in RerankModel:
+                return
+            return RerankModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], base_url=model_config["api_base"])
+
+        if llm_type == LLMType.IMAGE2TEXT.value:
+            if model_config["llm_factory"] not in CvModel:
+                return
+            return CvModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], lang, base_url=model_config["api_base"], **kwargs)
+
+        if llm_type == LLMType.CHAT.value:
+            if model_config["llm_factory"] not in ChatModel:
+                return
+            return ChatModel[model_config["llm_factory"]](model_config["api_key"], model_config["llm_name"], base_url=model_config["api_base"], **kwargs)
+
+        if llm_type == LLMType.SPEECH2TEXT:
+            if model_config["llm_factory"] not in Seq2txtModel:
+                return
+            return Seq2txtModel[model_config["llm_factory"]](key=model_config["api_key"], model_name=model_config["llm_name"], lang=lang, base_url=model_config["api_base"])
+        if llm_type == LLMType.TTS:
+            if model_config["llm_factory"] not in TTSModel:
+                return
+            return TTSModel[model_config["llm_factory"]](
+                model_config["api_key"],
+                model_config["llm_name"],
+                base_url=model_config["api_base"],
+            )
+
+    @classmethod
+    @DB.connection_context()
+    def increase_usage(cls, tenant_id, llm_type, used_tokens, llm_name=None):
+        e, tenant = TenantService.get_by_id(tenant_id)
+        if not e:
+            logging.error(f"Tenant not found: {tenant_id}")
+            return 0
+
+        llm_map = {
+            LLMType.EMBEDDING.value: tenant.embd_id if not llm_name else llm_name,
+            LLMType.SPEECH2TEXT.value: tenant.asr_id,
+            LLMType.IMAGE2TEXT.value: tenant.img2txt_id,
+            LLMType.CHAT.value: tenant.llm_id if not llm_name else llm_name,
+            LLMType.RERANK.value: tenant.rerank_id if not llm_name else llm_name,
+            LLMType.TTS.value: tenant.tts_id if not llm_name else llm_name,
+        }
+
+        mdlnm = llm_map.get(llm_type)
+        if mdlnm is None:
+            logging.error(f"LLM type error: {llm_type}")
+            return 0
+
+        llm_name, llm_factory = TenantLLMService.split_model_name_and_factory(mdlnm)
+
+        try:
+            num = (
+                cls.model.update(used_tokens=cls.model.used_tokens + used_tokens)
+                .where(cls.model.tenant_id == tenant_id, cls.model.llm_name == llm_name, cls.model.llm_factory == llm_factory if llm_factory else True)
+                .execute()
+            )
+        except Exception:
+            logging.exception("TenantLLMService.increase_usage got exception,Failed to update used_tokens for tenant_id=%s, llm_name=%s", tenant_id, llm_name)
+            return 0
+
+        return num
+
+    @classmethod
+    @DB.connection_context()
+    def get_openai_models(cls):
+        objs = cls.model.select().where((cls.model.llm_factory == "OpenAI"), ~(cls.model.llm_name == "text-embedding-3-small"), ~(cls.model.llm_name == "text-embedding-3-large")).dicts()
+        return list(objs)
+
+    @staticmethod
+    def llm_id2llm_type(llm_id: str) -> str | None:
+        from api.db.services.llm_service import LLMService
+        llm_id, *_ = TenantLLMService.split_model_name_and_factory(llm_id)
+        llm_factories = settings.FACTORY_LLM_INFOS
+        for llm_factory in llm_factories:
+            for llm in llm_factory["llm"]:
+                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
+
+        llm = TenantLLMService.get_or_none(llm_name=llm_id)
+        if llm:
+            return llm.model_type
+        for llm in TenantLLMService.query(llm_name=llm_id):
+            return llm.model_type
+
+
+class LLM4Tenant:
+    def __init__(self, tenant_id, llm_type, llm_name=None, lang="Chinese", **kwargs):
+        self.tenant_id = tenant_id
+        self.llm_type = llm_type
+        self.llm_name = llm_name
+        self.mdl = TenantLLMService.model_instance(tenant_id, llm_type, llm_name, lang=lang, **kwargs)
+        assert self.mdl, "Can't find model for {}/{}/{}".format(tenant_id, llm_type, llm_name)
+        model_config = TenantLLMService.get_model_config(tenant_id, llm_type, llm_name)
+        self.max_length = model_config.get("max_tokens", 8192)
+
+        self.is_tools = model_config.get("is_tools", False)
+        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
+                trace_id = self.langfuse.create_trace_id()
+                self.trace_context = {"trace_id": trace_id}

api/ragflow_server.py

@@ -33,7 +33,7 @@ import uuid
 
 from werkzeug.serving import run_simple
 from api import settings
-from api.apps import app
+from api.apps import app, smtp_mail_server
 from api.db.runtime_config import RuntimeConfig
 from api.db.services.document_service import DocumentService
 from api import utils
@@ -59,11 +59,14 @@ def update_progress():
             if redis_lock.acquire():
                 DocumentService.update_progress()
                 redis_lock.release()
-            stop_event.wait(6)
         except Exception:
             logging.exception("update_progress exception")
         finally:
-            redis_lock.release()
+            try:
+                redis_lock.release()
+            except Exception:
+                logging.exception("update_progress exception")
+            stop_event.wait(6)
 
 def signal_handler(sig, frame):
     logging.info("Received interrupt signal, shutting down...")
@@ -74,11 +77,11 @@ def signal_handler(sig, frame):
 
 if __name__ == '__main__':
     logging.info(r"""
-        ____   ___    ______ ______ __               
+        ____   ___    ______ ______ __
        / __ \ /   |  / ____// ____// /____  _      __
       / /_/ // /| | / / __ / /_   / // __ \| | /| / /
-     / _, _// ___ |/ /_/ // __/  / // /_/ /| |/ |/ / 
-    /_/ |_|/_/  |_|\____//_/    /_/ \____/ |__/|__/                             
+     / _, _// ___ |/ /_/ // __/  / // /_/ /| |/ |/ /
+    /_/ |_|/_/  |_|\____//_/    /_/ \____/ |__/|__/
 
     """)
     logging.info(
@@ -137,6 +140,18 @@ if __name__ == '__main__':
     else:
         threading.Timer(1.0, delayed_start_update_progress).start()
 
+    # init smtp server
+    if settings.SMTP_CONF:
+        app.config["MAIL_SERVER"] = settings.MAIL_SERVER
+        app.config["MAIL_PORT"] = settings.MAIL_PORT
+        app.config["MAIL_USE_SSL"] = settings.MAIL_USE_SSL
+        app.config["MAIL_USE_TLS"] = settings.MAIL_USE_TLS
+        app.config["MAIL_USERNAME"] = settings.MAIL_USERNAME
+        app.config["MAIL_PASSWORD"] = settings.MAIL_PASSWORD
+        app.config["MAIL_DEFAULT_SENDER"] = settings.MAIL_DEFAULT_SENDER
+        smtp_mail_server.init_app(app)
+
+
     # start http server
     try:
         logging.info("RAGFlow HTTP server start...")

api/settings.py

@@ -38,6 +38,11 @@ EMBEDDING_MDL = ""
 RERANK_MDL = ""
 ASR_MDL = ""
 IMAGE2TEXT_MDL = ""
+CHAT_CFG = ""
+EMBEDDING_CFG = ""
+RERANK_CFG = ""
+ASR_CFG = ""
+IMAGE2TEXT_CFG = ""
 API_KEY = None
 PARSERS = None
 HOST_IP = None
@@ -74,23 +79,32 @@ STRONG_TEST_COUNT = int(os.environ.get("STRONG_TEST_COUNT", "8"))
 
 BUILTIN_EMBEDDING_MODELS = ["BAAI/bge-large-zh-v1.5@BAAI", "maidalun1020/bce-embedding-base_v1@Youdao"]
 
+SMTP_CONF = None
+MAIL_SERVER = ""
+MAIL_PORT = 000
+MAIL_USE_SSL= True
+MAIL_USE_TLS = False
+MAIL_USERNAME = ""
+MAIL_PASSWORD = ""
+MAIL_DEFAULT_SENDER = ()
+MAIL_FRONTEND_URL = ""
+
+
 def get_or_create_secret_key():
     secret_key = os.environ.get("RAGFLOW_SECRET_KEY")
     if secret_key and len(secret_key) >= 32:
         return secret_key
-    
+
     # Check if there's a configured secret key
     configured_key = get_base_config(RAG_FLOW_SERVICE_NAME, {}).get("secret_key")
     if configured_key and configured_key != str(date.today()) and len(configured_key) >= 32:
         return configured_key
-    
+
     # Generate a new secure key and warn about it
     import logging
+
     new_key = secrets.token_hex(32)
-    logging.warning(
-        "SECURITY WARNING: Using auto-generated SECRET_KEY. "
-        f"Generated key: {new_key}"
-    )
+    logging.warning(f"SECURITY WARNING: Using auto-generated SECRET_KEY. Generated key: {new_key}")
     return new_key
 
 
@@ -99,10 +113,10 @@ def init_settings():
     LIGHTEN = int(os.environ.get("LIGHTEN", "0"))
     DATABASE_TYPE = os.getenv("DB_TYPE", "mysql")
     DATABASE = decrypt_database_config(name=DATABASE_TYPE)
-    LLM = get_base_config("user_default_llm", {})
-    LLM_DEFAULT_MODELS = LLM.get("default_models", {})
-    LLM_FACTORY = LLM.get("factory")
-    LLM_BASE_URL = LLM.get("base_url")
+    LLM = get_base_config("user_default_llm", {}) or {}
+    LLM_DEFAULT_MODELS = LLM.get("default_models", {}) or {}
+    LLM_FACTORY = LLM.get("factory", "") or ""
+    LLM_BASE_URL = LLM.get("base_url", "") or ""
     try:
         REGISTER_ENABLED = int(os.environ.get("REGISTER_ENABLED", "1"))
     except Exception:
@@ -115,29 +129,34 @@ def init_settings():
         FACTORY_LLM_INFOS = []
 
     global CHAT_MDL, EMBEDDING_MDL, RERANK_MDL, ASR_MDL, IMAGE2TEXT_MDL
+    global CHAT_CFG, EMBEDDING_CFG, RERANK_CFG, ASR_CFG, IMAGE2TEXT_CFG
     if not LIGHTEN:
         EMBEDDING_MDL = BUILTIN_EMBEDDING_MODELS[0]
 
-    if LLM_DEFAULT_MODELS:
-        CHAT_MDL = LLM_DEFAULT_MODELS.get("chat_model", CHAT_MDL)
-        EMBEDDING_MDL = LLM_DEFAULT_MODELS.get("embedding_model", EMBEDDING_MDL)
-        RERANK_MDL = LLM_DEFAULT_MODELS.get("rerank_model", RERANK_MDL)
-        ASR_MDL = LLM_DEFAULT_MODELS.get("asr_model", ASR_MDL)
-        IMAGE2TEXT_MDL = LLM_DEFAULT_MODELS.get("image2text_model", IMAGE2TEXT_MDL)
-
-        # factory can be specified in the config name with "@". LLM_FACTORY will be used if not specified
-        CHAT_MDL = CHAT_MDL + (f"@{LLM_FACTORY}" if "@" not in CHAT_MDL and CHAT_MDL != "" else "")
-        EMBEDDING_MDL = EMBEDDING_MDL + (f"@{LLM_FACTORY}" if "@" not in EMBEDDING_MDL and EMBEDDING_MDL != "" else "")
-        RERANK_MDL = RERANK_MDL + (f"@{LLM_FACTORY}" if "@" not in RERANK_MDL and RERANK_MDL != "" else "")
-        ASR_MDL = ASR_MDL + (f"@{LLM_FACTORY}" if "@" not in ASR_MDL and ASR_MDL != "" else "")
-        IMAGE2TEXT_MDL = IMAGE2TEXT_MDL + (f"@{LLM_FACTORY}" if "@" not in IMAGE2TEXT_MDL and IMAGE2TEXT_MDL != "" else "")
-
     global API_KEY, PARSERS, HOST_IP, HOST_PORT, SECRET_KEY
     API_KEY = LLM.get("api_key")
     PARSERS = LLM.get(
         "parsers", "naive:General,qa:Q&A,resume:Resume,manual:Manual,table:Table,paper:Paper,book:Book,laws:Laws,presentation:Presentation,picture:Picture,one:One,audio:Audio,email:Email,tag:Tag"
     )
 
+    chat_entry = _parse_model_entry(LLM_DEFAULT_MODELS.get("chat_model", CHAT_MDL))
+    embedding_entry = _parse_model_entry(LLM_DEFAULT_MODELS.get("embedding_model", EMBEDDING_MDL))
+    rerank_entry = _parse_model_entry(LLM_DEFAULT_MODELS.get("rerank_model", RERANK_MDL))
+    asr_entry = _parse_model_entry(LLM_DEFAULT_MODELS.get("asr_model", ASR_MDL))
+    image2text_entry = _parse_model_entry(LLM_DEFAULT_MODELS.get("image2text_model", IMAGE2TEXT_MDL))
+
+    CHAT_CFG = _resolve_per_model_config(chat_entry, LLM_FACTORY, API_KEY, LLM_BASE_URL)
+    EMBEDDING_CFG = _resolve_per_model_config(embedding_entry, LLM_FACTORY, API_KEY, LLM_BASE_URL)
+    RERANK_CFG = _resolve_per_model_config(rerank_entry, LLM_FACTORY, API_KEY, LLM_BASE_URL)
+    ASR_CFG = _resolve_per_model_config(asr_entry, LLM_FACTORY, API_KEY, LLM_BASE_URL)
+    IMAGE2TEXT_CFG = _resolve_per_model_config(image2text_entry, LLM_FACTORY, API_KEY, LLM_BASE_URL)
+
+    CHAT_MDL = CHAT_CFG.get("model", "") or ""
+    EMBEDDING_MDL = EMBEDDING_CFG.get("model", "") or ""
+    RERANK_MDL = RERANK_CFG.get("model", "") or ""
+    ASR_MDL = ASR_CFG.get("model", "") or ""
+    IMAGE2TEXT_MDL = IMAGE2TEXT_CFG.get("model", "") or ""
+
     HOST_IP = get_base_config(RAG_FLOW_SERVICE_NAME, {}).get("host", "127.0.0.1")
     HOST_PORT = get_base_config(RAG_FLOW_SERVICE_NAME, {}).get("http_port")
 
@@ -170,12 +189,28 @@ def init_settings():
 
     retrievaler = search.Dealer(docStoreConn)
     from graphrag import search as kg_search
+
     kg_retrievaler = kg_search.KGSearch(docStoreConn)
 
     if int(os.environ.get("SANDBOX_ENABLED", "0")):
         global SANDBOX_HOST
         SANDBOX_HOST = os.environ.get("SANDBOX_HOST", "sandbox-executor-manager")
 
+    global SMTP_CONF, MAIL_SERVER, MAIL_PORT, MAIL_USE_SSL, MAIL_USE_TLS
+    global MAIL_USERNAME, MAIL_PASSWORD, MAIL_DEFAULT_SENDER, MAIL_FRONTEND_URL
+    SMTP_CONF = get_base_config("smtp", {})
+
+    MAIL_SERVER = SMTP_CONF.get("mail_server", "")
+    MAIL_PORT = SMTP_CONF.get("mail_port", 000)
+    MAIL_USE_SSL = SMTP_CONF.get("mail_use_ssl", True)
+    MAIL_USE_TLS = SMTP_CONF.get("mail_use_tls", False)
+    MAIL_USERNAME = SMTP_CONF.get("mail_username", "")
+    MAIL_PASSWORD = SMTP_CONF.get("mail_password", "")
+    mail_default_sender = SMTP_CONF.get("mail_default_sender", [])
+    if mail_default_sender and len(mail_default_sender) >= 2:
+        MAIL_DEFAULT_SENDER = (mail_default_sender[0], mail_default_sender[1])
+    MAIL_FRONTEND_URL = SMTP_CONF.get("mail_frontend_url", "")
+
 
 class CustomEnum(Enum):
     @classmethod
@@ -210,3 +245,34 @@ class RetCode(IntEnum, CustomEnum):
     SERVER_ERROR = 500
     FORBIDDEN = 403
     NOT_FOUND = 404
+
+
+def _parse_model_entry(entry):
+    if isinstance(entry, str):
+        return {"name": entry, "factory": None, "api_key": None, "base_url": None}
+    if isinstance(entry, dict):
+        name = entry.get("name") or entry.get("model") or ""
+        return {
+            "name": name,
+            "factory": entry.get("factory"),
+            "api_key": entry.get("api_key"),
+            "base_url": entry.get("base_url"),
+        }
+    return {"name": "", "factory": None, "api_key": None, "base_url": None}
+
+
+def _resolve_per_model_config(entry_dict, backup_factory, backup_api_key, backup_base_url):
+    name = (entry_dict.get("name") or "").strip()
+    m_factory = entry_dict.get("factory") or backup_factory or ""
+    m_api_key = entry_dict.get("api_key") or backup_api_key or ""
+    m_base_url = entry_dict.get("base_url") or backup_base_url or ""
+
+    if name and "@" not in name and m_factory:
+        name = f"{name}@{m_factory}"
+
+    return {
+        "model": name,
+        "factory": m_factory,
+        "api_key": m_api_key,
+        "base_url": m_base_url,
+    }

api/utils/api_utils.py

@@ -48,7 +48,8 @@ from werkzeug.http import HTTP_STATUS_CODES
 from api import settings
 from api.constants import REQUEST_MAX_WAIT_SEC, REQUEST_WAIT_SEC
 from api.db.db_models import APIToken
-from api.db.services.llm_service import LLMService, TenantLLMService
+from api.db.services.llm_service import LLMService
+from api.db.services.tenant_llm_service import TenantLLMService
 from api.utils import CustomJSONEncoder, get_uuid, json_dumps
 from rag.utils.mcp_tool_call_conn import MCPToolCallSession, close_multiple_mcp_toolcall_sessions
 

api/utils/web_utils.py

@@ -21,6 +21,9 @@ import re
 import socket
 from urllib.parse import urlparse
 
+from api.apps import smtp_mail_server
+from flask_mail import Message
+from flask import render_template_string
 from selenium import webdriver
 from selenium.common.exceptions import TimeoutException
 from selenium.webdriver.chrome.options import Options
@@ -31,6 +34,7 @@ from selenium.webdriver.support.ui import WebDriverWait
 from webdriver_manager.chrome import ChromeDriverManager
 
 
+
 CONTENT_TYPE_MAP = {
     # Office
     "docx": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
@@ -172,3 +176,26 @@ def get_float(req: dict, key: str, default: float | int = 10.0) -> float:
         return parsed if parsed > 0 else default
     except (TypeError, ValueError):
         return default
+
+
+INVITE_EMAIL_TMPL = """
+<p>Hi {{email}},</p>
+<p>{{inviter}} has invited you to join their team (ID: {{tenant_id}}).</p>
+<p>Click the link below to complete your registration:<br>
+<a href="{{invite_url}}">{{invite_url}}</a></p>
+<p>If you did not request this, please ignore this email.</p>
+"""
+
+def send_invite_email(to_email, invite_url, tenant_id, inviter):
+    from api.apps import  app
+    with app.app_context():
+        msg = Message(subject="RAGFlow Invitation",
+                      recipients=[to_email])
+        msg.html = render_template_string(
+            INVITE_EMAIL_TMPL,
+            email=to_email,
+            invite_url=invite_url,
+            tenant_id=tenant_id,
+            inviter=inviter,
+        )
+        smtp_mail_server.send(msg)

conf/llm_factories.json

@@ -1146,60 +1146,35 @@
                     "llm_name": "gemini-2.5-flash",
                     "tags": "LLM,CHAT,1024K,IMAGE2TEXT",
                     "max_tokens": 1048576,
-                    "model_type": "image2text",
+                    "model_type": "chat",
                     "is_tools": true
                 },
                 {
                     "llm_name": "gemini-2.5-pro",
                     "tags": "LLM,CHAT,IMAGE2TEXT,1024K",
                     "max_tokens": 1048576,
-                    "model_type": "image2text",
+                    "model_type": "chat",
                     "is_tools": true
                 },
                 {
-                    "llm_name": "gemini-2.5-flash-preview-05-20",
+                    "llm_name": "gemini-2.5-flash-lite",
                     "tags": "LLM,CHAT,1024K,IMAGE2TEXT",
                     "max_tokens": 1048576,
-                    "model_type": "image2text",
+                    "model_type": "chat",
                     "is_tools": true
                 },
                 {
-                    "llm_name": "gemini-2.0-flash-001",
-                    "tags": "LLM,CHAT,1024K",
-                    "max_tokens": 1048576,
-                    "model_type": "image2text",
-                    "is_tools": true
-                },
-                {
-                    "llm_name": "gemini-2.0-flash-thinking-exp-01-21",
+                    "llm_name": "gemini-2.0-flash",
                     "tags": "LLM,CHAT,1024K",
                     "max_tokens": 1048576,
                     "model_type": "chat",
                     "is_tools": true
                 },
                 {
-                    "llm_name": "gemini-1.5-flash",
-                    "tags": "LLM,IMAGE2TEXT,1024K",
+                    "llm_name": "gemini-2.0-flash-lite",
+                    "tags": "LLM,CHAT,1024K",
                     "max_tokens": 1048576,
-                    "model_type": "image2text"
-                },
-                {
-                    "llm_name": "gemini-2.5-pro-preview-05-06",
-                    "tags": "LLM,IMAGE2TEXT,1024K",
-                    "max_tokens": 1048576,
-                    "model_type": "image2text"
-                },
-                {
-                    "llm_name": "gemini-1.5-pro",
-                    "tags": "LLM,IMAGE2TEXT,2048K",
-                    "max_tokens": 2097152,
-                    "model_type": "image2text"
-                },
-                {
-                    "llm_name": "gemini-1.5-flash-8b",
-                    "tags": "LLM,IMAGE2TEXT,1024K",
-                    "max_tokens": 1048576,
-                    "model_type": "image2text",
+                    "model_type": "chat",
                     "is_tools": true
                 },
                 {

conf/service_conf.yaml

@@ -64,9 +64,21 @@ redis:
 #   config:
 #     oss_table: 'opendal_storage'
 # user_default_llm:
-#   factory: 'Tongyi-Qianwen'
-#   api_key: 'sk-xxxxxxxxxxxxx'
-#   base_url: ''
+#   factory: 'BAAI'
+#   api_key: 'backup'
+#   base_url: 'backup_base_url'
+#   default_models:
+#     chat_model:
+#       name: 'qwen2.5-7b-instruct'
+#       factory: 'xxxx'
+#       api_key: 'xxxx'
+#       base_url: 'https://api.xx.com'
+#     embedding_model:
+#       name: 'bge-m3'
+#     rerank_model: 'bge-reranker-v2'
+#     asr_model:
+#       model: 'whisper-large-v3' # alias of name
+#     image2text_model: ''
 # oauth:
 #   oauth2:
 #     display_name: "OAuth2"
@@ -101,3 +113,14 @@ redis:
 #   switch: false
 #   component: false
 #   dataset: false
+# smtp:
+#   mail_server: ""
+#   mail_port: 465
+#   mail_use_ssl: true
+#   mail_use_tls: false
+#   mail_username: ""
+#   mail_password: ""
+#   mail_default_sender:
+#     - "RAGFlow" # display name
+#     - "" # sender email address
+#   mail_frontend_url: "https://your-frontend.example.com"

deepdoc/parser/excel_parser.py

@@ -12,6 +12,7 @@
 #
 
 import logging
+import re
 import sys
 from io import BytesIO
 
@@ -20,6 +21,8 @@ from openpyxl import Workbook, load_workbook
 
 from rag.nlp import find_codec
 
+# copied from `/openpyxl/cell/cell.py`
+ILLEGAL_CHARACTERS_RE = re.compile(r'[\000-\010]|[\013-\014]|[\016-\037]')
 
 class RAGFlowExcelParser:
 
@@ -50,13 +53,29 @@ class RAGFlowExcelParser:
             logging.info(f"openpyxl load error: {e}, try pandas instead")
             try:
                 file_like_object.seek(0)
-                df = pd.read_excel(file_like_object)
-                return RAGFlowExcelParser._dataframe_to_workbook(df)
+                try:
+                    df = pd.read_excel(file_like_object)
+                    return RAGFlowExcelParser._dataframe_to_workbook(df)
+                except Exception as ex:
+                    logging.info(f"pandas with default engine load error: {ex}, try calamine instead")
+                    file_like_object.seek(0)
+                    df = pd.read_excel(file_like_object, engine='calamine')
+                    return RAGFlowExcelParser._dataframe_to_workbook(df)
             except Exception as e_pandas:
                 raise Exception(f"pandas.read_excel error: {e_pandas}, original openpyxl error: {e}")
 
+    @staticmethod
+    def _clean_dataframe(df: pd.DataFrame):
+        def clean_string(s):
+            if isinstance(s, str):
+                return ILLEGAL_CHARACTERS_RE.sub(" ", s)
+            return s
+
+        return df.apply(lambda col: col.map(clean_string))
+
     @staticmethod
     def _dataframe_to_workbook(df):
+        df = RAGFlowExcelParser._clean_dataframe(df)
         wb = Workbook()
         ws = wb.active
         ws.title = "Data"
@@ -71,9 +90,17 @@ class RAGFlowExcelParser:
         return wb
 
     def html(self, fnm, chunk_rows=256):
+        from html import escape
+
         file_like_object = BytesIO(fnm) if not isinstance(fnm, str) else fnm
         wb = RAGFlowExcelParser._load_excel_to_workbook(file_like_object)
         tb_chunks = []
+
+        def _fmt(v):
+            if v is None:
+                return ""
+            return str(v).strip()
+
         for sheetname in wb.sheetnames:
             ws = wb[sheetname]
             rows = list(ws.rows)
@@ -82,7 +109,7 @@ class RAGFlowExcelParser:
 
             tb_rows_0 = "<tr>"
             for t in list(rows[0]):
-                tb_rows_0 += f"<th>{t.value}</th>"
+                tb_rows_0 += f"<th>{escape(_fmt(t.value))}</th>"
             tb_rows_0 += "</tr>"
 
             for chunk_i in range((len(rows) - 1) // chunk_rows + 1):
@@ -90,7 +117,7 @@ class RAGFlowExcelParser:
                 tb += f"<table><caption>{sheetname}</caption>"
                 tb += tb_rows_0
                 for r in list(
-                    rows[1 + chunk_i * chunk_rows: 1 + (chunk_i + 1) * chunk_rows]
+                    rows[1 + chunk_i * chunk_rows: min(1 + (chunk_i + 1) * chunk_rows, len(rows))]
                 ):
                     tb += "<tr>"
                     for i, c in enumerate(r):

docs/guides/agent/sandbox_quickstart.md

@@ -63,7 +63,7 @@ docker build -t sandbox-executor-manager:latest ./executor_manager
 3. Add the following entry to your /etc/hosts file to resolve the executor manager service:
 
 ```bash
-127.0.0.1 sandbox-executor-manager
+127.0.0.1 es01 infinity mysql minio redis sandbox-executor-manager
 ```
 
 4. Start the RAGFlow service as usual.

docs/references/python_api_reference.md

@@ -5,7 +5,7 @@ slug: /python_api_reference
 
 # Python API
 
-A complete reference for RAGFlow's Python APIs. Before proceeding, please ensure you [have your RAGFlow API key ready for authentication](../guides/models/llm_api_key_setup.md).
+A complete reference for RAGFlow's Python APIs. Before proceeding, please ensure you [have your RAGFlow API key ready for authentication](../develop/acquire_ragflow_api_key.md).
 
 :::tip NOTE
 Run the following command to download the Python SDK:

docs/release_notes.md

@@ -9,8 +9,8 @@ Key features, improvements and bug fixes in the latest releases.
 
 :::info
 Each RAGFlow release is available in two editions:
-- **Slim edition**: excludes built-in embedding models and is identified by a **-slim** suffix added to the version name. Example: `infiniflow/ragflow:v0.19.1-slim`
-- **Full edition**: includes built-in embedding models and has no suffix added to the version name. Example: `infiniflow/ragflow:v0.19.1`
+- **Slim edition**: excludes built-in embedding models and is identified by a **-slim** suffix added to the version name. Example: `infiniflow/ragflow:v0.20.1-slim`
+- **Full edition**: includes built-in embedding models and has no suffix added to the version name. Example: `infiniflow/ragflow:v0.20.1`
 :::
 
 :::danger IMPORTANT
@@ -22,6 +22,36 @@ 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.
 :::
 
+## v0.20.2 (Ongoing🔨)
+
+Released on August ??, 2025.
+
+### Improvements
+
+- Revamps the user interface for the **Datasets**, **Chat**, and **Search** pages.  
+- Search: Supports creating search apps tailored to various business scenarios
+- Chat: Supports comparing answer performance of up to three chat model settings on a single **Chat** page.
+- Agent:  
+  - Implements a toggle in the **Agent** component to enable or disable citation.  
+  - Introduces a drag-and-drop method for creating components.  
+- Documentation: Corrects inaccuracies in the API reference.
+
+### New Agent templates
+
+- Report Agent: A template for generating summary reports in internal question-answering scenarios, supporting the display of tables and formulae.  [#9427](https://github.com/infiniflow/ragflow/pull/9427)
+
+### Fixed issues
+
+- Predefined opening greeting in the **Agent** component was missing during conversations.  
+- An automatic line break issue in the prompt editor.  
+- A memory leak issue caused by PyPDF. [#9469](https://github.com/infiniflow/ragflow/pull/9469)
+
+### API changes
+
+#### Deprecated
+
+[Create session with agent](./references/http_api_reference.md#create-session-with-agent)
+
 ## v0.20.1
 
 Released on August 8, 2025.
@@ -33,10 +63,10 @@ Released on August 8, 2025.
 
 ### Added Models
 
-- ChatGPT 5
+- GPT-5
 - Claude 4.1
 
-### New agent Templates (both workflow and agentic)
+### New agent templates (both workflow and agentic)
 
 - SQL Assistant Workflow: Empowers non-technical teams (e.g., operations, product) to independently query business data.
 - Choose Your Knowledge Base Workflow: Lets users select a knowledge base to query during conversations. [#9325](https://github.com/infiniflow/ragflow/pull/9325)
@@ -182,7 +212,7 @@ From this release onwards, if you still see RAGFlow's responses being cut short
 
 - Unable to add models via Ollama/Xinference, an issue introduced in v0.17.1.
 
-### Related APIs
+### API changes
 
 #### HTTP APIs
 
@@ -243,7 +273,7 @@ The following is a screenshot of a conversation that integrates Deep Research:
 
 ![Image](https://github.com/user-attachments/assets/165b88ff-1f5d-4fb8-90e2-c836b25e32e9)
 
-### Related APIs
+### API changes
 
 #### HTTP APIs
 
@@ -318,7 +348,7 @@ This release fixes the following issues:
 - Using the **Table** parsing method results in information loss.
 - Miscellaneous API issues.
 
-### Related APIs
+### API changes
 
 #### HTTP APIs
 
@@ -354,7 +384,7 @@ Released on December 18, 2024.
 - Upgrades the Document Layout Analysis model in DeepDoc.
 - Significantly enhances the retrieval performance when using [Infinity](https://github.com/infiniflow/infinity) as document engine.
 
-### Related APIs
+### API changes
 
 #### HTTP APIs
 
@@ -411,7 +441,7 @@ This approach eliminates the need to manually update **service_config.yaml** aft
 Ensure that you [upgrade **both** your code **and** Docker image to this release](https://ragflow.io/docs/dev/upgrade_ragflow#upgrade-ragflow-to-the-most-recent-officially-published-release) before trying this new approach.
 :::
 
-### Related APIs
+### API changes
 
 #### HTTP APIs
 
@@ -570,7 +600,7 @@ While we also test RAGFlow on ARM64 platforms, we do not maintain RAGFlow Docker
 If you are on an ARM platform, follow [this guide](./develop/build_docker_image.mdx) to build a RAGFlow Docker image.
 :::
 
-### Related APIs
+### API changes
 
 #### HTTP API
 
@@ -591,7 +621,7 @@ Released on May 21, 2024.
 - Supports monitoring of system components, including Elasticsearch, MySQL, Redis, and MinIO.
 - Supports disabling **Layout Recognition** in the GENERAL chunking method to reduce file chunking time.
 
-### Related APIs
+### API changes
 
 #### HTTP API
 

graphrag/entity_resolution.py

@@ -106,7 +106,7 @@ class EntityResolution(Extractor):
             nonlocal remain_candidates_to_resolve, callback
             async with semaphore:
                 try:
-                    with trio.move_on_after(180) as cancel_scope:
+                    with trio.move_on_after(280) as cancel_scope:
                         await self._resolve_candidate(candidate_batch, result_set, result_lock)
                         remain_candidates_to_resolve = remain_candidates_to_resolve - len(candidate_batch[1])
                         callback(msg=f"Resolved {len(candidate_batch[1])} pairs, {remain_candidates_to_resolve} are remained to resolve. ")
@@ -169,7 +169,7 @@ class EntityResolution(Extractor):
         logging.info(f"Created resolution prompt {len(text)} bytes for {len(candidate_resolution_i[1])} entity pairs of type {candidate_resolution_i[0]}")
         async with chat_limiter:
             try:
-                with trio.move_on_after(120) as cancel_scope:
+                with trio.move_on_after(240) as cancel_scope:
                     response = await trio.to_thread.run_sync(self._chat, text, [{"role": "user", "content": "Output:"}], {})
                 if cancel_scope.cancelled_caught:
                     logging.warning("_resolve_candidate._chat timeout, skipping...")

graphrag/general/community_reports_extractor.py

@@ -92,7 +92,7 @@ class CommunityReportsExtractor(Extractor):
             text = perform_variable_replacements(self._extraction_prompt, variables=prompt_variables)
             async with chat_limiter:
                 try:
-                    with trio.move_on_after(80) as cancel_scope:
+                    with trio.move_on_after(180) as cancel_scope:
                         response = await trio.to_thread.run_sync( self._chat, text, [{"role": "user", "content": "Output:"}], {})
                     if cancel_scope.cancelled_caught:
                         logging.warning("extract_community_report._chat timeout, skipping...")

graphrag/general/index.py

@@ -57,20 +57,22 @@ async def run_graphrag(
     ):
         chunks.append(d["content_with_weight"])
 
-    subgraph = await generate_subgraph(
-        LightKGExt
-        if "method" not in row["kb_parser_config"].get("graphrag", {}) or row["kb_parser_config"]["graphrag"]["method"] != "general"
-        else GeneralKGExt,
-        tenant_id,
-        kb_id,
-        doc_id,
-        chunks,
-        language,
-        row["kb_parser_config"]["graphrag"].get("entity_types", []),
-        chat_model,
-        embedding_model,
-        callback,
-    )
+    with trio.fail_after(max(120, len(chunks)*120)):
+        subgraph = await generate_subgraph(
+            LightKGExt
+            if "method" not in row["kb_parser_config"].get("graphrag", {}) or row["kb_parser_config"]["graphrag"]["method"] != "general"
+            else GeneralKGExt,
+            tenant_id,
+            kb_id,
+            doc_id,
+            chunks,
+            language,
+            row["kb_parser_config"]["graphrag"].get("entity_types", []),
+            chat_model,
+            embedding_model,
+            callback,
+        )
+
     if not subgraph:
         return
 
@@ -125,7 +127,6 @@ async def run_graphrag(
     return
 
 
-@timeout(60*60, 1)
 async def generate_subgraph(
     extractor: Extractor,
     tenant_id: str,

helm/templates/elasticsearch.yaml

@@ -44,9 +44,21 @@ spec:
         checksum/config-es: {{ include (print $.Template.BasePath "/elasticsearch-config.yaml") . | sha256sum }}
         checksum/config-env: {{ include (print $.Template.BasePath "/env.yaml") . | sha256sum }}
     spec:
+      {{- if or .Values.imagePullSecrets .Values.elasticsearch.image.pullSecrets }}
+      imagePullSecrets:
+        {{- with .Values.imagePullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.elasticsearch.image.pullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
       initContainers:
       - name: fix-data-volume-permissions
-        image: alpine
+        image: {{ .Values.elasticsearch.initContainers.alpine.repository }}:{{ .Values.elasticsearch.initContainers.alpine.tag }}
+        {{- with .Values.elasticsearch.initContainers.alpine.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         command:
         - sh
         - -c
@@ -55,14 +67,20 @@ spec:
           - mountPath: /usr/share/elasticsearch/data
             name: es-data
       - name: sysctl
-        image: busybox
+        image: {{ .Values.elasticsearch.initContainers.busybox.repository }}:{{ .Values.elasticsearch.initContainers.busybox.tag }}
+        {{- with .Values.elasticsearch.initContainers.busybox.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         securityContext:
           privileged: true
           runAsUser: 0
         command: ["sysctl", "-w", "vm.max_map_count=262144"]
       containers:
       - name: elasticsearch
-        image: elasticsearch:{{ .Values.env.STACK_VERSION }}
+        image: {{ .Values.elasticsearch.image.repository }}:{{ .Values.elasticsearch.image.tag }}
+        {{- with .Values.elasticsearch.image.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         envFrom:
           - secretRef:
               name: {{ include "ragflow.fullname" . }}-env-config

helm/templates/infinity.yaml

@@ -43,9 +43,21 @@ spec:
       annotations:
         checksum/config: {{ include (print $.Template.BasePath "/env.yaml") . | sha256sum }}
     spec:
+      {{- if or .Values.imagePullSecrets .Values.infinity.image.pullSecrets }}
+      imagePullSecrets:
+        {{- with .Values.imagePullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.infinity.image.pullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
       containers:
       - name: infinity
         image: {{ .Values.infinity.image.repository }}:{{ .Values.infinity.image.tag }}
+        {{- with .Values.infinity.image.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         envFrom:
           - secretRef:
               name: {{ include "ragflow.fullname" . }}-env-config

helm/templates/minio.yaml

@@ -43,9 +43,21 @@ spec:
         {{- include "ragflow.labels" . | nindent 8 }}
         app.kubernetes.io/component: minio
     spec:
+      {{- if or .Values.imagePullSecrets .Values.minio.image.pullSecrets }}
+      imagePullSecrets:
+        {{- with .Values.imagePullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.minio.image.pullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
       containers:
       - name: minio
         image: {{ .Values.minio.image.repository }}:{{ .Values.minio.image.tag }}
+        {{- with .Values.minio.image.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         envFrom:
           - secretRef:
               name: {{ include "ragflow.fullname" . }}-env-config

helm/templates/mysql.yaml

@@ -44,9 +44,21 @@ spec:
         checksum/config-mysql: {{ include (print $.Template.BasePath "/mysql-config.yaml") . | sha256sum }}
         checksum/config-env: {{ include (print $.Template.BasePath "/env.yaml") . | sha256sum }}
     spec:
+      {{- if or .Values.imagePullSecrets .Values.mysql.image.pullSecrets }}
+      imagePullSecrets:
+        {{- with .Values.imagePullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.mysql.image.pullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
       containers:
       - name: mysql
         image: {{ .Values.mysql.image.repository }}:{{ .Values.mysql.image.tag }}
+        {{- with .Values.mysql.image.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         envFrom:
           - secretRef:
               name: {{ include "ragflow.fullname" . }}-env-config

helm/templates/opensearch.yaml

@@ -44,9 +44,21 @@ spec:
         checksum/config-opensearch: {{ include (print $.Template.BasePath "/opensearch-config.yaml") . | sha256sum }}
         checksum/config-env: {{ include (print $.Template.BasePath "/env.yaml") . | sha256sum }}
     spec:
+      {{- if or .Values.imagePullSecrets .Values.opensearch.image.pullSecrets }}
+      imagePullSecrets:
+        {{- with .Values.imagePullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.opensearch.image.pullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
       initContainers:
       - name: fix-data-volume-permissions
-        image: alpine
+        image: {{ .Values.opensearch.initContainers.alpine.repository }}:{{ .Values.opensearch.initContainers.alpine.tag }}
+        {{- with .Values.opensearch.initContainers.alpine.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         command:
         - sh
         - -c
@@ -55,7 +67,10 @@ spec:
           - mountPath: /usr/share/opensearch/data
             name: opensearch-data
       - name: sysctl
-        image: busybox
+        image: {{ .Values.opensearch.initContainers.busybox.repository }}:{{ .Values.opensearch.initContainers.busybox.tag }}
+        {{- with .Values.opensearch.initContainers.busybox.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         securityContext:
           privileged: true
           runAsUser: 0
@@ -63,6 +78,9 @@ spec:
       containers:
       - name: opensearch
         image: {{ .Values.opensearch.image.repository }}:{{ .Values.opensearch.image.tag }}
+        {{- with .Values.opensearch.image.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         envFrom:
           - secretRef:
               name: {{ include "ragflow.fullname" . }}-env-config

helm/templates/ragflow.yaml

@@ -25,9 +25,21 @@ spec:
         checksum/config-env: {{ include (print $.Template.BasePath "/env.yaml") . | sha256sum }}
         checksum/config-ragflow: {{ include (print $.Template.BasePath "/ragflow_config.yaml") . | sha256sum }}
     spec:
+      {{- if or .Values.imagePullSecrets .Values.ragflow.image.pullSecrets }}
+      imagePullSecrets:
+        {{- with .Values.imagePullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.ragflow.image.pullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
       containers:
       - name: ragflow
-        image: {{ .Values.env.RAGFLOW_IMAGE }}
+        image: {{ .Values.ragflow.image.repository }}:{{ .Values.ragflow.image.tag }}
+        {{- with .Values.ragflow.image.pullPolicy }}
+        imagePullPolicy: {{ . }}
+        {{- end }}
         ports:
           - containerPort: 80
             name: http

helm/templates/redis.yaml

@@ -40,10 +40,22 @@ spec:
       annotations:
         checksum/config-env: {{ include (print $.Template.BasePath "/env.yaml") . | sha256sum }}
     spec:
+      {{- if or .Values.imagePullSecrets .Values.redis.image.pullSecrets }}
+      imagePullSecrets:
+        {{- with .Values.imagePullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+        {{- with .Values.redis.image.pullSecrets }}
+        {{- toYaml . | nindent 8 }}
+        {{- end }}
+      {{- end }}
       terminationGracePeriodSeconds: 60
       containers:
         - name: redis
           image: {{ .Values.redis.image.repository }}:{{ .Values.redis.image.tag }}
+          {{- with .Values.redis.image.pullPolicy }}
+          imagePullPolicy: {{ . }}
+          {{- end }}
           command:
             - "sh"
             - "-c"

helm/values.yaml

@@ -1,4 +1,8 @@
 # Based on docker compose .env file
+
+# Global image pull secrets configuration
+imagePullSecrets: []
+
 env:
   # The type of doc engine to use.
   # Available options:
@@ -32,31 +36,6 @@ env:
   # The password for Redis
   REDIS_PASSWORD: infini_rag_flow_helm
 
-  # The RAGFlow Docker image to download.
-  # Defaults to the v0.20.1-slim edition, which is the RAGFlow Docker image without embedding models.
-  RAGFLOW_IMAGE: infiniflow/ragflow:v0.20.1-slim
-  #
-  # To download the RAGFlow Docker image with embedding models, uncomment the following line instead:
-  # RAGFLOW_IMAGE: infiniflow/ragflow:v0.20.1
-  #
-  # The Docker image of the v0.20.1 edition includes:
-  # - Built-in embedding models:
-  #   - BAAI/bge-large-zh-v1.5
-  #   - BAAI/bge-reranker-v2-m3
-  #   - maidalun1020/bce-embedding-base_v1
-  #   - maidalun1020/bce-reranker-base_v1
-  # - Embedding models that will be downloaded once you select them in the RAGFlow UI:
-  #   - BAAI/bge-base-en-v1.5
-  #   - BAAI/bge-large-en-v1.5
-  #   - BAAI/bge-small-en-v1.5
-  #   - BAAI/bge-small-zh-v1.5
-  #   - jinaai/jina-embeddings-v2-base-en
-  #   - jinaai/jina-embeddings-v2-small-en
-  #   - nomic-ai/nomic-embed-text-v1.5
-  #   - sentence-transformers/all-MiniLM-L6-v2
-  #
-  #
-
   # The local time zone.
   TIMEZONE: "Asia/Shanghai"
 
@@ -75,7 +54,11 @@ env:
   EMBEDDING_BATCH_SIZE: 16
 
 ragflow:
-
+  image:
+    repository: infiniflow/ragflow
+    tag: v0.20.1-slim
+    pullPolicy: IfNotPresent
+    pullSecrets: []
   # Optional service configuration overrides
   # to be written to local.service_conf.yaml
   # inside the RAGFlow container
@@ -114,6 +97,8 @@ infinity:
   image:
     repository: infiniflow/infinity
     tag: v0.6.0-dev5
+    pullPolicy: IfNotPresent
+    pullSecrets: []
   storage:
     className:
     capacity: 5Gi
@@ -124,6 +109,20 @@ infinity:
     type: ClusterIP
 
 elasticsearch:
+  image:
+    repository: elasticsearch
+    tag: "8.11.3"
+    pullPolicy: IfNotPresent
+    pullSecrets: []
+  initContainers:
+    alpine:
+      repository: alpine
+      tag: latest
+      pullPolicy: IfNotPresent
+    busybox:
+      repository: busybox
+      tag: latest
+      pullPolicy: IfNotPresent
   storage:
     className:
     capacity: 20Gi
@@ -140,6 +139,17 @@ opensearch:
   image:
     repository: opensearchproject/opensearch
     tag: 2.19.1
+    pullPolicy: IfNotPresent
+    pullSecrets: []
+  initContainers:
+    alpine:
+      repository: alpine
+      tag: latest
+      pullPolicy: IfNotPresent
+    busybox:
+      repository: busybox
+      tag: latest
+      pullPolicy: IfNotPresent
   storage:
     className:
     capacity: 20Gi
@@ -156,6 +166,8 @@ minio:
   image:
     repository: quay.io/minio/minio
     tag: RELEASE.2023-12-20T01-00-02Z
+    pullPolicy: IfNotPresent
+    pullSecrets: []
   storage:
     className:
     capacity: 5Gi
@@ -169,6 +181,8 @@ mysql:
   image:
     repository: mysql
     tag: 8.0.39
+    pullPolicy: IfNotPresent
+    pullSecrets: []
   storage:
     className:
     capacity: 5Gi
@@ -182,6 +196,8 @@ redis:
   image:
     repository: valkey/valkey
     tag: 8
+    pullPolicy: IfNotPresent
+    pullSecrets: []
   storage:
     className:
     capacity: 5Gi

mcp/server/server.py

@@ -180,7 +180,7 @@ async def list_tools(*, connector) -> list[types.Tool]:
     return [
         types.Tool(
             name="ragflow_retrieval",
-            description="Retrieve relevant chunks from the RAGFlow retrieve interface based on the question, using the specified dataset_ids and optionally document_ids. Below is the list of all available datasets, including their descriptions and IDs. If you're unsure which datasets are relevant to the question, simply pass all dataset IDs to the function."
+            description="Retrieve relevant chunks from the RAGFlow retrieve interface based on the question. You can optionally specify dataset_ids to search only specific datasets, or omit dataset_ids entirely to search across ALL available datasets. You can also optionally specify document_ids to search within specific documents. When dataset_ids is not provided or is empty, the system will automatically search across all available datasets. Below is the list of all available datasets, including their descriptions and IDs:"
             + dataset_description,
             inputSchema={
                 "type": "object",
@@ -188,14 +188,16 @@ async def list_tools(*, connector) -> list[types.Tool]:
                     "dataset_ids": {
                         "type": "array",
                         "items": {"type": "string"},
+                        "description": "Optional array of dataset IDs to search. If not provided or empty, all datasets will be searched."
                     },
                     "document_ids": {
                         "type": "array",
                         "items": {"type": "string"},
+                        "description": "Optional array of document IDs to search within."
                     },
-                    "question": {"type": "string"},
+                    "question": {"type": "string", "description": "The question or query to search for."},
                 },
-                "required": ["dataset_ids", "question"],
+                "required": ["question"],
             },
         ),
     ]
@@ -206,8 +208,26 @@ async def list_tools(*, connector) -> list[types.Tool]:
 async def call_tool(name: str, arguments: dict, *, connector) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:
     if name == "ragflow_retrieval":
         document_ids = arguments.get("document_ids", [])
+        dataset_ids = arguments.get("dataset_ids", [])
+        
+        # If no dataset_ids provided or empty list, get all available dataset IDs
+        if not dataset_ids:
+            dataset_list_str = connector.list_datasets()
+            dataset_ids = []
+            
+            # Parse the dataset list to extract IDs
+            if dataset_list_str:
+                for line in dataset_list_str.strip().split('\n'):
+                    if line.strip():
+                        try:
+                            dataset_info = json.loads(line.strip())
+                            dataset_ids.append(dataset_info["id"])
+                        except (json.JSONDecodeError, KeyError):
+                            # Skip malformed lines
+                            continue
+        
         return connector.retrieval(
-            dataset_ids=arguments["dataset_ids"],
+            dataset_ids=dataset_ids,
             document_ids=document_ids,
             question=arguments["question"],
         )

pyproject.toml

@@ -24,7 +24,7 @@ dependencies = [
     "chardet==5.2.0",
     "cn2an==0.5.22",
     "cohere==5.6.2",
-    "Crawl4AI==0.3.8",
+    "Crawl4AI>=0.3.8",
     "dashscope==1.20.11",
     "deepl==1.18.0",
     "demjson3==3.0.6",
@@ -43,7 +43,7 @@ dependencies = [
     "groq==0.9.0",
     "hanziconv==0.3.2",
     "html-text==0.6.2",
-    "httpx==0.27.0",
+    "httpx[socks]==0.27.2",
     "huggingface-hub>=0.25.0,<0.26.0",
     "infinity-sdk==0.6.0-dev4",
     "infinity-emb>=0.0.66,<0.0.67",
@@ -58,7 +58,7 @@ dependencies = [
     "ollama==0.2.1",
     "onnxruntime==1.19.2; sys_platform == 'darwin' or platform_machine != 'x86_64'",
     "onnxruntime-gpu==1.19.2; sys_platform != 'darwin' and platform_machine == 'x86_64'",
-    "openai==1.45.0",
+    "openai>=1.45.0",
     "opencv-python==4.10.0.84",
     "opencv-python-headless==4.10.0.84",
     "openpyxl>=3.1.0,<4.0.0",
@@ -73,7 +73,7 @@ dependencies = [
     "pyclipper==1.3.0.post5",
     "pycryptodomex==3.20.0",
     "pymysql>=1.1.1,<2.0.0",
-    "pypdf>=5.0.0,<6.0.0",
+    "pypdf==6.0.0",
     "python-dotenv==1.0.1",
     "python-dateutil==2.8.2",
     "python-pptx>=1.0.2,<2.0.0",
@@ -128,6 +128,9 @@ dependencies = [
     "opensearch-py==2.7.1",
     "pluginlib==0.9.4",
     "click>=8.1.8",
+    "python-calamine>=0.4.0",
+    "litellm>=1.74.15.post1",
+    "flask-mail>=0.10.0",
 ]
 
 [project.optional-dependencies]

rag/app/naive.py

@@ -22,6 +22,8 @@ from timeit import default_timer as timer
 
 from docx import Document
 from docx.image.exceptions import InvalidImageStreamError, UnexpectedEndOfFileError, UnrecognizedImageError
+from docx.opc.pkgreader import _SerializedRelationships, _SerializedRelationship
+from docx.opc.oxml import parse_xml
 from markdown import markdown
 from PIL import Image
 from tika import parser
@@ -47,8 +49,8 @@ class Docx(DocxParser):
         if not embed:
             return None
         embed = embed[0]
-        related_part = document.part.related_parts[embed]
         try:
+            related_part = document.part.related_parts[embed]
             image_blob = related_part.image.blob
         except UnrecognizedImageError:
             logging.info("Unrecognized image format. Skipping image.")
@@ -62,6 +64,9 @@ class Docx(DocxParser):
         except UnicodeDecodeError:
             logging.info("The recognized image stream appears to be corrupted. Skipping image.")
             return None
+        except Exception:
+            logging.info("The recognized image stream appears to be corrupted. Skipping image.")
+            return None
         try:
             image = Image.open(BytesIO(image_blob)).convert('RGB')
             return image
@@ -226,17 +231,20 @@ class Docx(DocxParser):
             for r in tb.rows:
                 html += "<tr>"
                 i = 0
-                while i < len(r.cells):
-                    span = 1
-                    c = r.cells[i]
-                    for j in range(i + 1, len(r.cells)):
-                        if c.text == r.cells[j].text:
-                            span += 1
-                            i = j
-                        else:
-                            break
-                    i += 1
-                    html += f"<td>{c.text}</td>" if span == 1 else f"<td colspan='{span}'>{c.text}</td>"
+                try:
+                    while i < len(r.cells):
+                        span = 1
+                        c = r.cells[i]
+                        for j in range(i + 1, len(r.cells)):
+                            if c.text == r.cells[j].text:
+                                span += 1
+                                i = j
+                            else:
+                                break
+                        i += 1
+                        html += f"<td>{c.text}</td>" if span == 1 else f"<td colspan='{span}'>{c.text}</td>"
+                except Exception as e:
+                    logging.warning(f"Error parsing table, ignore: {e}")
                 html += "</tr>"
             html += "</table>"
             tbls.append(((None, html), ""))
@@ -357,6 +365,20 @@ class Markdown(MarkdownParser):
             tbls.append(((None, markdown(table, extensions=['markdown.extensions.tables'])), ""))
         return sections, tbls
 
+def load_from_xml_v2(baseURI, rels_item_xml):
+    """
+    Return |_SerializedRelationships| instance loaded with the
+    relationships contained in *rels_item_xml*. Returns an empty
+    collection if *rels_item_xml* is |None|.
+    """
+    srels = _SerializedRelationships()
+    if rels_item_xml is not None:
+        rels_elm = parse_xml(rels_item_xml)
+        for rel_elm in rels_elm.Relationship_lst:
+            if rel_elm.target_ref in ('../NULL', 'NULL'):
+                continue
+            srels._srels.append(_SerializedRelationship(baseURI, rel_elm))
+    return srels
 
 def chunk(filename, binary=None, from_page=0, to_page=100000,
           lang="Chinese", callback=None, **kwargs):
@@ -388,6 +410,8 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
         except Exception:
             vision_model = None
 
+        # fix "There is no item named 'word/NULL' in the archive", referring to https://github.com/python-openxml/python-docx/issues/1105#issuecomment-1298075246
+        _SerializedRelationships.load_from_xml = load_from_xml_v2
         sections, tables = Docx()(filename, binary)
 
         if vision_model:
@@ -466,6 +490,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             sections = [(_, "") for _ in excel_parser.html(binary, 12) if _]
         else:
             sections = [(_, "") for _ in excel_parser(binary) if _]
+        parser_config["chunk_token_num"] = 12800
 
     elif re.search(r"\.(txt|py|js|java|c|cpp|h|php|go|ts|sh|cs|kt|sql)$", filename, re.IGNORECASE):
         callback(0.1, "Start to parse.")

rag/app/table.py

@@ -40,7 +40,6 @@ class Excel(ExcelParser):
         total = 0
         for sheetname in wb.sheetnames:
             total += len(list(wb[sheetname].rows))
-
         res, fails, done = [], [], 0
         rn = 0
         for sheetname in wb.sheetnames:
@@ -48,31 +47,204 @@ class Excel(ExcelParser):
             rows = list(ws.rows)
             if not rows:
                 continue
-            headers = [cell.value for cell in rows[0]]
-            missed = set([i for i, h in enumerate(headers) if h is None])
-            headers = [cell.value for i, cell in enumerate(rows[0]) if i not in missed]
+            headers, header_rows = self._parse_headers(ws, rows)
             if not headers:
                 continue
             data = []
-            for i, r in enumerate(rows[1:]):
+            for i, r in enumerate(rows[header_rows:]):
                 rn += 1
                 if rn - 1 < from_page:
                     continue
                 if rn - 1 >= to_page:
                     break
-                row = [cell.value for ii, cell in enumerate(r) if ii not in missed]
-                if len(row) != len(headers):
+                row_data = self._extract_row_data(ws, r, header_rows + i, len(headers))
+                if row_data is None:
                     fails.append(str(i))
                     continue
-                data.append(row)
+                if self._is_empty_row(row_data):
+                    continue
+                data.append(row_data)
                 done += 1
-            if np.array(data).size == 0:
+            if len(data) == 0:
                 continue
-            res.append(pd.DataFrame(np.array(data), columns=headers))
-
+            df = pd.DataFrame(data, columns=headers)
+            res.append(df)
         callback(0.3, ("Extract records: {}~{}".format(from_page + 1, min(to_page, from_page + rn)) + (f"{len(fails)} failure, line: %s..." % (",".join(fails[:3])) if fails else "")))
         return res
 
+    def _parse_headers(self, ws, rows):
+        if len(rows) == 0:
+            return [], 0
+        has_complex_structure = self._has_complex_header_structure(ws, rows)
+        if has_complex_structure:
+            return self._parse_multi_level_headers(ws, rows)
+        else:
+            return self._parse_simple_headers(rows)
+
+    def _has_complex_header_structure(self, ws, rows):
+        if len(rows) < 1:
+            return False
+        merged_ranges = list(ws.merged_cells.ranges)
+        # 检查前两行是否涉及合并单元格
+        for rng in merged_ranges:
+            if rng.min_row <= 2:  # 只要合并区域涉及第1或第2行
+                return True
+        return False
+
+    def _row_looks_like_header(self, row):
+        header_like_cells = 0
+        data_like_cells = 0
+        non_empty_cells = 0
+        for cell in row:
+            if cell.value is not None:
+                non_empty_cells += 1
+                val = str(cell.value).strip()
+                if self._looks_like_header(val):
+                    header_like_cells += 1
+                elif self._looks_like_data(val):
+                    data_like_cells += 1
+        if non_empty_cells == 0:
+            return False
+        return header_like_cells >= data_like_cells
+
+    def _parse_simple_headers(self, rows):
+        if not rows:
+            return [], 0
+        header_row = rows[0]
+        headers = []
+        for cell in header_row:
+            if cell.value is not None:
+                header_value = str(cell.value).strip()
+                if header_value:
+                    headers.append(header_value)
+            else:
+                pass
+        final_headers = []
+        for i, cell in enumerate(header_row):
+            if cell.value is not None:
+                header_value = str(cell.value).strip()
+                if header_value:
+                    final_headers.append(header_value)
+                else:
+                    final_headers.append(f"Column_{i + 1}")
+            else:
+                final_headers.append(f"Column_{i + 1}")
+        return final_headers, 1
+
+    def _parse_multi_level_headers(self, ws, rows):
+        if len(rows) < 2:
+            return [], 0
+        header_rows = self._detect_header_rows(rows)
+        if header_rows == 1:
+            return self._parse_simple_headers(rows)
+        else:
+            return self._build_hierarchical_headers(ws, rows, header_rows), header_rows
+
+    def _detect_header_rows(self, rows):
+        if len(rows) < 2:
+            return 1
+        header_rows = 1
+        max_check_rows = min(5, len(rows))
+        for i in range(1, max_check_rows):
+            row = rows[i]
+            if self._row_looks_like_header(row):
+                header_rows = i + 1
+            else:
+                break
+        return header_rows
+
+    def _looks_like_header(self, value):
+        if len(value) < 1:
+            return False
+        if any(ord(c) > 127 for c in value):
+            return True
+        if len([c for c in value if c.isalpha()]) >= 2:
+            return True
+        if any(c in value for c in ["(", ")", "：", ":", "（", "）", "_", "-"]):
+            return True
+        return False
+
+    def _looks_like_data(self, value):
+        if len(value) == 1 and value.upper() in ["Y", "N", "M", "X", "/", "-"]:
+            return True
+        if value.replace(".", "").replace("-", "").replace(",", "").isdigit():
+            return True
+        if value.startswith("0x") and len(value) <= 10:
+            return True
+        return False
+
+    def _build_hierarchical_headers(self, ws, rows, header_rows):
+        headers = []
+        max_col = max(len(row) for row in rows[:header_rows]) if header_rows > 0 else 0
+        merged_ranges = list(ws.merged_cells.ranges)
+        for col_idx in range(max_col):
+            header_parts = []
+            for row_idx in range(header_rows):
+                if col_idx < len(rows[row_idx]):
+                    cell_value = rows[row_idx][col_idx].value
+                    merged_value = self._get_merged_cell_value(ws, row_idx + 1, col_idx + 1, merged_ranges)
+                    if merged_value is not None:
+                        cell_value = merged_value
+                    if cell_value is not None:
+                        cell_value = str(cell_value).strip()
+                        if cell_value and cell_value not in header_parts and self._is_valid_header_part(cell_value):
+                            header_parts.append(cell_value)
+            if header_parts:
+                header = "-".join(header_parts)
+                headers.append(header)
+            else:
+                headers.append(f"Column_{col_idx + 1}")
+        final_headers = [h for h in headers if h and h != "-"]
+        return final_headers
+
+    def _is_valid_header_part(self, value):
+        if len(value) == 1 and value.upper() in ["Y", "N", "M", "X"]:
+            return False
+        if value.replace(".", "").replace("-", "").replace(",", "").isdigit():
+            return False
+        if value in ["/", "-", "+", "*", "="]:
+            return False
+        return True
+
+    def _get_merged_cell_value(self, ws, row, col, merged_ranges):
+        for merged_range in merged_ranges:
+            if merged_range.min_row <= row <= merged_range.max_row and merged_range.min_col <= col <= merged_range.max_col:
+                return ws.cell(merged_range.min_row, merged_range.min_col).value
+        return None
+
+    def _extract_row_data(self, ws, row, absolute_row_idx, expected_cols):
+        row_data = []
+        merged_ranges = list(ws.merged_cells.ranges)
+        actual_row_num = absolute_row_idx + 1
+        for col_idx in range(expected_cols):
+            cell_value = None
+            actual_col_num = col_idx + 1
+            try:
+                cell_value = ws.cell(row=actual_row_num, column=actual_col_num).value
+            except ValueError:
+                if col_idx < len(row):
+                    cell_value = row[col_idx].value
+            if cell_value is None:
+                merged_value = self._get_merged_cell_value(ws, actual_row_num, actual_col_num, merged_ranges)
+                if merged_value is not None:
+                    cell_value = merged_value
+                else:
+                    cell_value = self._get_inherited_value(ws, actual_row_num, actual_col_num, merged_ranges)
+            row_data.append(cell_value)
+        return row_data
+
+    def _get_inherited_value(self, ws, row, col, merged_ranges):
+        for merged_range in merged_ranges:
+            if merged_range.min_row <= row <= merged_range.max_row and merged_range.min_col <= col <= merged_range.max_col:
+                return ws.cell(merged_range.min_row, merged_range.min_col).value
+        return None
+
+    def _is_empty_row(self, row_data):
+        for val in row_data:
+            if val is not None and str(val).strip() != "":
+                return False
+        return True
+
 
 def trans_datatime(s):
     try:

rag/llm/__init__.py

@@ -19,6 +19,48 @@
 import importlib
 import inspect
 
+from strenum import StrEnum
+
+
+class SupportedLiteLLMProvider(StrEnum):
+    Tongyi_Qianwen = "Tongyi-Qianwen"
+    Dashscope = "Dashscope"
+    Bedrock = "Bedrock"
+    Moonshot = "Moonshot"
+    xAI = "xAI"
+    DeepInfra = "DeepInfra"
+    Groq = "Groq"
+    Cohere = "Cohere"
+    Gemini = "Gemini"
+    DeepSeek = "DeepSeek"
+    Nvidia = "NVIDIA"
+    TogetherAI = "TogetherAI"
+    Anthropic = "Anthropic"
+
+
+FACTORY_DEFAULT_BASE_URL = {
+    SupportedLiteLLMProvider.Tongyi_Qianwen: "https://dashscope.aliyuncs.com/compatible-mode/v1",
+    SupportedLiteLLMProvider.Dashscope: "https://dashscope.aliyuncs.com/compatible-mode/v1",
+    SupportedLiteLLMProvider.Moonshot: "https://api.moonshot.cn/v1",
+}
+
+
+LITELLM_PROVIDER_PREFIX = {
+    SupportedLiteLLMProvider.Tongyi_Qianwen: "dashscope/",
+    SupportedLiteLLMProvider.Dashscope: "dashscope/",
+    SupportedLiteLLMProvider.Bedrock: "bedrock/",
+    SupportedLiteLLMProvider.Moonshot: "moonshot/",
+    SupportedLiteLLMProvider.xAI: "xai/",
+    SupportedLiteLLMProvider.DeepInfra: "deepinfra/",
+    SupportedLiteLLMProvider.Groq: "groq/",
+    SupportedLiteLLMProvider.Cohere: "",  # don't need a prefix
+    SupportedLiteLLMProvider.Gemini: "gemini/",
+    SupportedLiteLLMProvider.DeepSeek: "deepseek/",
+    SupportedLiteLLMProvider.Nvidia: "nvidia_nim/",
+    SupportedLiteLLMProvider.TogetherAI: "together_ai/",
+    SupportedLiteLLMProvider.Anthropic: "",  # don't need a prefix
+}
+
 ChatModel = globals().get("ChatModel", {})
 CvModel = globals().get("CvModel", {})
 EmbeddingModel = globals().get("EmbeddingModel", {})
@@ -26,6 +68,7 @@ RerankModel = globals().get("RerankModel", {})
 Seq2txtModel = globals().get("Seq2txtModel", {})
 TTSModel = globals().get("TTSModel", {})
 
+
 MODULE_MAPPING = {
     "chat_model": ChatModel,
     "cv_model": CvModel,
@@ -42,20 +85,30 @@ for module_name, mapping_dict in MODULE_MAPPING.items():
     module = importlib.import_module(full_module_name)
 
     base_class = None
+    lite_llm_base_class = None
     for name, obj in inspect.getmembers(module):
-        if inspect.isclass(obj) and name == "Base":
-            base_class = obj
-            break
-    if base_class is None:
-        continue
+        if inspect.isclass(obj):
+            if name == "Base":
+                base_class = obj
+            elif name == "LiteLLMBase":
+                lite_llm_base_class = obj
+                assert hasattr(obj, "_FACTORY_NAME"), "LiteLLMbase should have _FACTORY_NAME field."
+                if hasattr(obj, "_FACTORY_NAME"):
+                    if isinstance(obj._FACTORY_NAME, list):
+                        for factory_name in obj._FACTORY_NAME:
+                            mapping_dict[factory_name] = obj
+                    else:
+                        mapping_dict[obj._FACTORY_NAME] = obj
+
+    if base_class is not None:
+        for _, obj in inspect.getmembers(module):
+            if inspect.isclass(obj) and issubclass(obj, base_class) and obj is not base_class and hasattr(obj, "_FACTORY_NAME"):
+                if isinstance(obj._FACTORY_NAME, list):
+                    for factory_name in obj._FACTORY_NAME:
+                        mapping_dict[factory_name] = obj
+                else:
+                    mapping_dict[obj._FACTORY_NAME] = obj
 
-    for _, obj in inspect.getmembers(module):
-        if inspect.isclass(obj) and issubclass(obj, base_class) and obj is not base_class and hasattr(obj, "_FACTORY_NAME"):
-            if isinstance(obj._FACTORY_NAME, list):
-                for factory_name in obj._FACTORY_NAME:
-                    mapping_dict[factory_name] = obj
-            else:
-                mapping_dict[obj._FACTORY_NAME] = obj
 
 __all__ = [
     "ChatModel",

rag/llm/cv_model.py

@@ -68,7 +68,7 @@ class Base(ABC):
             pmpt.append({
                 "type": "image_url",
                 "image_url": {
-                    "url": f"data:image/jpeg;base64,{img}" if img[:4] != "data" else img
+                    "url": img if isinstance(img, str) and img.startswith("data:") else f"data:image/png;base64,{img}"
                 }
             })
         return pmpt
@@ -109,16 +109,33 @@ class Base(ABC):
 
     @staticmethod
     def image2base64(image):
+        # Return a data URL with the correct MIME to avoid provider mismatches
         if isinstance(image, bytes):
-            return base64.b64encode(image).decode("utf-8")
+            # Best-effort magic number sniffing
+            mime = "image/png"
+            if len(image) >= 2 and image[0] == 0xFF and image[1] == 0xD8:
+                mime = "image/jpeg"
+            b64 = base64.b64encode(image).decode("utf-8")
+            return f"data:{mime};base64,{b64}"
         if isinstance(image, BytesIO):
-            return base64.b64encode(image.getvalue()).decode("utf-8")
+            data = image.getvalue()
+            mime = "image/png"
+            if len(data) >= 2 and data[0] == 0xFF and data[1] == 0xD8:
+                mime = "image/jpeg"
+            b64 = base64.b64encode(data).decode("utf-8")
+            return f"data:{mime};base64,{b64}"
         buffered = BytesIO()
+        fmt = "JPEG"
         try:
             image.save(buffered, format="JPEG")
         except Exception:
+            buffered = BytesIO()  # reset buffer before saving PNG
             image.save(buffered, format="PNG")
-        return base64.b64encode(buffered.getvalue()).decode("utf-8")
+            fmt = "PNG"
+        data = buffered.getvalue()
+        b64 = base64.b64encode(data).decode("utf-8")
+        mime = f"image/{fmt.lower()}"
+        return f"data:{mime};base64,{b64}"
 
     def prompt(self, b64):
         return [
@@ -372,6 +389,16 @@ class OllamaCV(Base):
         self.keep_alive = kwargs.get("ollama_keep_alive", int(os.environ.get("OLLAMA_KEEP_ALIVE", -1)))
         Base.__init__(self, **kwargs)
 
+
+    def _clean_img(self, img):
+        if not isinstance(img, str):
+            return img
+
+        #remove the header like "data/*;base64,"
+        if img.startswith("data:") and ";base64," in img:
+            img = img.split(";base64,")[1]
+        return img
+
     def _clean_conf(self, gen_conf):
         options = {}
         if "temperature" in gen_conf:
@@ -390,9 +417,12 @@ class OllamaCV(Base):
             hist.insert(0, {"role": "system", "content": system})
         if not images:
             return hist
+        temp_images = []
+        for img in images:
+            temp_images.append(self._clean_img(img))
         for his in hist:
             if his["role"] == "user":
-                his["images"] = images
+                his["images"] = temp_images
                 break
         return hist
 
@@ -509,24 +539,24 @@ class GeminiCV(Base):
         return res.text, res.usage_metadata.total_token_count
 
     def chat(self, system, history, gen_conf, images=[]):
-        from transformers import GenerationConfig
+        generation_config = dict(temperature=gen_conf.get("temperature", 0.3), top_p=gen_conf.get("top_p", 0.7))
         try:
             response = self.model.generate_content(
                 self._form_history(system, history, images),
-                generation_config=GenerationConfig(temperature=gen_conf.get("temperature", 0.3), top_p=gen_conf.get("top_p", 0.7)))
+                generation_config=generation_config)
             ans = response.text
             return ans, response.usage_metadata.total_token_count
         except Exception as e:
             return "**ERROR**: " + str(e), 0
 
     def chat_streamly(self, system, history, gen_conf, images=[]):
-        from transformers import GenerationConfig
         ans = ""
         response = None
         try:
+            generation_config = dict(temperature=gen_conf.get("temperature", 0.3), top_p=gen_conf.get("top_p", 0.7))
             response = self.model.generate_content(
                 self._form_history(system, history, images),
-                generation_config=GenerationConfig(temperature=gen_conf.get("temperature", 0.3), top_p=gen_conf.get("top_p", 0.7)),
+                generation_config=generation_config,
                 stream=True,
             )
 
@@ -542,7 +572,7 @@ class GeminiCV(Base):
             yield response.usage_metadata.total_token_count
         else:
             yield 0
-            
+
 
 class NvidiaCV(Base):
     _FACTORY_NAME = "NVIDIA"
@@ -661,8 +691,8 @@ class AnthropicCV(Base):
                         "type": "image",
                         "source": {
                             "type": "base64",
-                            "media_type": "image/jpeg" if img[:4] != "data" else img.split(":")[1].split(";")[0],
-                            "data": img if img[:4] != "data" else img.split(",")[1]
+                            "media_type": (img.split(":")[1].split(";")[0] if isinstance(img, str) and img[:4] == "data" else "image/png"),
+                            "data": (img.split(",")[1] if isinstance(img, str) and img[:4] == "data" else img)
                         },
                     }
             )

rag/llm/rerank_model.py

@@ -100,7 +100,7 @@ class DefaultRerank(Base):
         old_dynamic_batch_size = self._dynamic_batch_size
         if max_batch_size is not None:
             self._dynamic_batch_size = max_batch_size
-        res = np.array([], dtype=float)
+        res = np.array(len(pairs), dtype=float)
         i = 0
         while i < len(pairs):
             cur_i = i
@@ -111,7 +111,7 @@ class DefaultRerank(Base):
                 try:
                     # call subclass implemented batch processing calculation
                     batch_scores = self._compute_batch_scores(pairs[i : i + current_batch])
-                    res = np.append(res, batch_scores)
+                    res[i : i + current_batch] = batch_scores
                     i += current_batch
                     self._dynamic_batch_size = min(self._dynamic_batch_size * 2, 8)
                     break
@@ -125,8 +125,8 @@ class DefaultRerank(Base):
                         raise
             if retry_count >= max_retries:
                 raise RuntimeError("max retry times, still cannot process batch, please check your GPU memory")
-            self.torch_empty_cache()
-
+            
+        self.torch_empty_cache()
         self._dynamic_batch_size = old_dynamic_batch_size
         return np.array(res)
 
@@ -268,7 +268,7 @@ class LocalAIRerank(Base):
         max_rank = np.max(rank)
 
         # Avoid division by zero if all ranks are identical
-        if max_rank - min_rank != 0:
+        if not np.isclose(min_rank, max_rank, atol=1e-3):
             rank = (rank - min_rank) / (max_rank - min_rank)
         else:
             rank = np.zeros_like(rank)
@@ -482,9 +482,10 @@ class VoyageRerank(Base):
         self.model_name = model_name
 
     def similarity(self, query: str, texts: list):
-        rank = np.zeros(len(texts), dtype=float)
         if not texts:
-            return rank, 0
+            return np.array([]), 0
+        rank = np.zeros(len(texts), dtype=float)
+
         res = self.client.rerank(query=query, documents=texts, model=self.model_name, top_k=len(texts))
         try:
             for r in res.results:

rag/nlp/__init__.py

@@ -611,10 +611,6 @@ def naive_merge_with_images(texts, images, chunk_token_num=128, delimiter="\n。
                 if re.match(f"^{dels}$", sub_sec):
                     continue
                 add_chunk(sub_sec, image)
-        
-        for img in images:
-            if isinstance(img, Image.Image):
-                img.close()
 
     return cks, result_images
 

rag/nlp/search.py

@@ -383,8 +383,6 @@ class Dealer:
         vector_column = f"q_{dim}_vec"
         zero_vector = [0.0] * dim
         sim_np = np.array(sim)
-        if doc_ids:
-            similarity_threshold = 0
         filtered_count = (sim_np >= similarity_threshold).sum()    
         ranks["total"] = int(filtered_count) # Convert from np.int64 to Python int otherwise JSON serializable error
         for i in idx:

rag/prompts/analyze_task_user.md

@@ -4,6 +4,9 @@ Task: {{ task }}
 
 Context: {{ context }}
 
+**Agent Prompt**
+{{ agent_prompt }}
+
 **Analysis Requirements:**
 1. Is it just a small talk? (If yes, no further plan or analysis is needed)
 2. What is the core objective of the task?

rag/prompts/ask_summary.md

@@ -0,0 +1,14 @@
+Role: You're a smart assistant. Your name is Miss R.
+Task: Summarize the information from knowledge bases and answer user's question.
+Requirements and restriction:
+  - DO NOT make things up, especially for numbers.
+  - If the information from knowledge is irrelevant with user's question, JUST SAY: Sorry, no relevant information provided.
+  - Answer with markdown format text.
+  - Answer in language of user's question.
+  - DO NOT make things up, especially for numbers.
+
+### Information from knowledge bases
+
+{{ knowledge }}
+
+The above is information from knowledge bases.

rag/prompts/citation_prompt.md

@@ -105,4 +105,5 @@ REMEMBER:
 - Cite FACTS, not opinions or transitions
 - Each citation supports the ENTIRE sentence
 - When in doubt, ask: "Would a fact-checker need to verify this?"
-- Place citations at sentence end, before punctuation
+- Place citations at sentence end, before punctuation
+- Format likes this is FORBIDDEN: [ID:0, ID:5, ID:...]. It MUST be seperated like, [ID:0][ID:5]...

rag/prompts/meta_filter.md

@@ -0,0 +1,53 @@
+You are a metadata filtering condition generator. Analyze the user's question and available document metadata to output a JSON array of filter objects. Follow these rules:
+
+1. **Metadata Structure**: 
+   - Metadata is provided as JSON where keys are attribute names (e.g., "color"), and values are objects mapping attribute values to document IDs.
+   - Example: 
+     {
+       "color": {"red": ["doc1"], "blue": ["doc2"]},
+       "listing_date": {"2025-07-11": ["doc1"], "2025-08-01": ["doc2"]}
+     }
+
+2. **Output Requirements**:
+   - Always output a JSON array of filter objects
+   - Each object must have:
+        "key": (metadata attribute name),
+        "value": (string value to compare),
+        "op": (operator from allowed list)
+
+3. **Operator Guide**:
+   - Use these operators only: ["contains", "not contains", "start with", "end with", "empty", "not empty", "=", "≠", ">", "<", "≥", "≤"]
+   - Date ranges: Break into two conditions (≥ start_date AND < next_month_start)
+   - Negations: Always use "≠" for exclusion terms ("not", "except", "exclude", "≠")
+   - Implicit logic: Derive unstated filters (e.g., "July" → [≥ YYYY-07-01, < YYYY-08-01])
+
+4. **Processing Steps**:
+   a) Identify ALL filterable attributes in the query (both explicit and implicit)
+   b) For dates:
+        - Infer missing year from current date if needed
+        - Always format dates as "YYYY-MM-DD"
+        - Convert ranges: [≥ start, < end]
+   c) For values: Match EXACTLY to metadata's value keys
+   d) Skip conditions if:
+        - Attribute doesn't exist in metadata
+        - Value has no match in metadata
+
+5. **Example**:
+   - User query: "上市日期七月份的有哪些商品，不要蓝色的"
+   - Metadata: { "color": {...}, "listing_date": {...} }
+   - Output: 
+        [
+          {"key": "listing_date", "value": "2025-07-01", "op": "≥"},
+          {"key": "listing_date", "value": "2025-08-01", "op": "<"},
+          {"key": "color", "value": "blue", "op": "≠"}
+        ]
+
+6. **Final Output**:
+   - ONLY output valid JSON array
+   - NO additional text/explanations
+
+**Current Task**:
+- Today's date: {{current_date}}
+- Available metadata keys: {{metadata_keys}}
+- User query: "{{user_question}}"
+

rag/prompts/prompts.py

@@ -149,6 +149,8 @@ NEXT_STEP = load_prompt("next_step")
 REFLECT = load_prompt("reflect")
 SUMMARY4MEMORY = load_prompt("summary4memory")
 RANK_MEMORY = load_prompt("rank_memory")
+META_FILTER = load_prompt("meta_filter")
+ASK_SUMMARY = load_prompt("ask_summary")
 
 PROMPT_JINJA_ENV = jinja2.Environment(autoescape=False, trim_blocks=True, lstrip_blocks=True)
 
@@ -196,7 +198,7 @@ def question_proposal(chat_mdl, content, topn=3):
 def full_question(tenant_id=None, llm_id=None, messages=[], language=None, chat_mdl=None):
     from api.db import LLMType
     from api.db.services.llm_service import LLMBundle
-    from api.db.services.llm_service import TenantLLMService
+    from api.db.services.tenant_llm_service import TenantLLMService
 
     if not chat_mdl:
         if TenantLLMService.llm_id2llm_type(llm_id) == "image2text":
@@ -230,7 +232,7 @@ def full_question(tenant_id=None, llm_id=None, messages=[], language=None, chat_
 def cross_languages(tenant_id, llm_id, query, languages=[]):
     from api.db import LLMType
     from api.db.services.llm_service import LLMBundle
-    from api.db.services.llm_service import TenantLLMService
+    from api.db.services.tenant_llm_service import TenantLLMService
 
     if llm_id and TenantLLMService.llm_id2llm_type(llm_id) == "image2text":
         chat_mdl = LLMBundle(tenant_id, LLMType.IMAGE2TEXT, llm_id)
@@ -335,13 +337,13 @@ def form_history(history, limit=-6):
     return context
 
 
-def analyze_task(chat_mdl, task_name, tools_description: list[dict]):
+def analyze_task(chat_mdl, prompt, task_name, tools_description: list[dict]):
     tools_desc = tool_schema(tools_description)
     context = ""
 
     template = PROMPT_JINJA_ENV.from_string(ANALYZE_TASK_USER)
-
-    kwd = chat_mdl.chat(ANALYZE_TASK_SYSTEM,[{"role": "user", "content": template.render(task=task_name, context=context, tools_desc=tools_desc)}], {})
+    context = template.render(task=task_name, context=context, agent_prompt=prompt, tools_desc=tools_desc)
+    kwd = chat_mdl.chat(ANALYZE_TASK_SYSTEM,[{"role": "user", "content": context}], {})
     if isinstance(kwd, tuple):
         kwd = kwd[0]
     kwd = re.sub(r"^.*</think>", "", kwd, flags=re.DOTALL)
@@ -413,3 +415,20 @@ def rank_memories(chat_mdl, goal:str, sub_goal:str, tool_call_summaries: list[st
     ans = chat_mdl.chat(msg[0]["content"], msg[1:], stop="<|stop|>")
     return re.sub(r"^.*</think>", "", ans, flags=re.DOTALL)
 
+
+def gen_meta_filter(chat_mdl, meta_data:dict, query: str) -> list:
+    sys_prompt = PROMPT_JINJA_ENV.from_string(META_FILTER).render(
+        current_date=datetime.datetime.today().strftime('%Y-%m-%d'),
+        metadata_keys=json.dumps(meta_data),
+        user_question=query
+    )
+    user_prompt = "Generate filters:"
+    ans = chat_mdl.chat(sys_prompt, [{"role": "user", "content": user_prompt}])
+    ans = re.sub(r"(^.*</think>|```json\n|```\n*$)", "", ans, flags=re.DOTALL)
+    try:
+        ans = json_repair.loads(ans)
+        assert isinstance(ans, list), ans
+        return ans
+    except Exception:
+        logging.exception(f"Loading json failure: {ans}")
+    return []

rag/prompts/related_question.md

@@ -0,0 +1,55 @@
+# Role
+You are an AI language model assistant tasked with generating **5-10 related questions** based on a user’s original query.
+These questions should help **expand the search query scope** and **improve search relevance**.
+
+---
+
+## Instructions
+
+**Input:**
+You are provided with a **user’s question**.
+
+**Output:**
+Generate **5-10 alternative questions** that are **related** to the original user question.
+These alternatives should help retrieve a **broader range of relevant documents** from a vector database.
+
+**Context:**
+Focus on **rephrasing** the original question in different ways, ensuring the alternative questions are **diverse but still connected** to the topic of the original query.
+Do **not** create overly obscure, irrelevant, or unrelated questions.
+
+**Fallback:**
+If you cannot generate any relevant alternatives, do **not** return any questions.
+
+---
+
+## Guidance
+
+1. Each alternative should be **unique** but still **relevant** to the original query.
+2. Keep the phrasing **clear, concise, and easy to understand**.
+3. Avoid overly technical jargon or specialized terms **unless directly relevant**.
+4. Ensure that each question **broadens** the search angle, **not narrows** it.
+
+---
+
+## Example
+
+**Original Question:**
+> What are the benefits of electric vehicles?
+
+**Alternative Questions:**
+1. How do electric vehicles impact the environment?
+2. What are the advantages of owning an electric car?
+3. What is the cost-effectiveness of electric vehicles?
+4. How do electric vehicles compare to traditional cars in terms of fuel efficiency?
+5. What are the environmental benefits of switching to electric cars?
+6. How do electric vehicles help reduce carbon emissions?
+7. Why are electric vehicles becoming more popular?
+8. What are the long-term savings of using electric vehicles?
+9. How do electric vehicles contribute to sustainability?
+10. What are the key benefits of electric vehicles for consumers?
+
+---
+
+## Reason
+Rephrasing the original query into multiple alternative questions helps the user explore **different aspects** of their search topic, improving the **quality of search results**.
+These questions guide the search engine to provide a **more comprehensive set** of relevant documents.

rag/svr/task_executor.py

@@ -302,9 +302,13 @@ async def build_chunks(task, progress_callback):
                     # If the image is in RGBA mode, convert it to RGB mode before saving it in JPEG format.
                     if d["image"].mode in ("RGBA", "P"):
                         converted_image = d["image"].convert("RGB")
-                        d["image"].close()  # Close original image
+                        #d["image"].close()  # Close original image
                         d["image"] = converted_image
-                    d["image"].save(output_buffer, format='JPEG')
+                    try:
+                        d["image"].save(output_buffer, format='JPEG')
+                    except OSError as e:
+                        logging.warning(
+                            "Saving image of chunk {}/{}/{} got exception, ignore: {}".format(task["location"], task["name"], d["id"], str(e)))
 
                 async with minio_limiter:
                     await trio.to_thread.run_sync(lambda: STORAGE_IMPL.put(task["kb_id"], d["id"], output_buffer.getvalue()))
@@ -440,7 +444,7 @@ 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)
+    @timeout(60)
     def batch_encode(txts):
         nonlocal mdl
         return mdl.encode([truncate(c, mdl.max_length-10) for c in txts])
@@ -516,7 +520,7 @@ async def run_raptor(row, chat_mdl, embd_mdl, vector_size, callback=None):
     return res, tk_count
 
 
-@timeout(60*60, 1)
+@timeout(60*60*2, 1)
 async def do_handle_task(task):
     task_id = task["id"]
     task_from_page = task["from_page"]

rag/utils/s3_conn.py

@@ -190,3 +190,16 @@ class RAGFlowS3:
                 self.__open__()
                 time.sleep(1)
         return
+
+    @use_default_bucket
+    def rm_bucket(self, bucket, *args, **kwargs):
+        for conn in self.conn:
+            try:
+                if not conn.bucket_exists(bucket):
+                    continue
+                for o in conn.list_objects_v2(Bucket=bucket):
+                    conn.delete_object(bucket, o.object_name)
+                conn.delete_bucket(Bucket=bucket)
+                return
+            except Exception as e:
+                logging.error(f"Fail rm {bucket}: " + str(e))

sdk/python/ragflow_sdk/modules/chat.py

@@ -24,7 +24,6 @@ class Chat(Base):
         self.id = ""
         self.name = "assistant"
         self.avatar = "path/to/avatar"
-        self.dataset_ids = ["kb1"]
         self.llm = Chat.LLM(rag, {})
         self.prompt = Chat.Prompt(rag, {})
         super().__init__(rag, res_dict)
@@ -48,7 +47,7 @@ class Chat(Base):
             self.variables = [{"key": "knowledge", "optional": True}]
             self.rerank_model = ""
             self.empty_response = None
-            self.opener = "Hi! I'm your assistant, what can I do for you?"
+            self.opener = "Hi! I'm your assistant. What can I do for you?"
             self.show_quote = True
             self.prompt = (
                 "You are an intelligent assistant. Please summarize the content of the knowledge base to answer the question. "

sdk/python/ragflow_sdk/modules/document.py

@@ -44,6 +44,7 @@ class Document(Base):
         self.process_duration = 0.0
         self.run = "0"
         self.status = "1"
+        self.meta_fields = {}
         for k in list(res_dict.keys()):
             if k not in self.__dict__:
                 res_dict.pop(k)

sdk/python/ragflow_sdk/ragflow.py

@@ -143,7 +143,7 @@ class RAGFlow:
                 },
             )
             if prompt.opener is None:
-                prompt.opener = "Hi! I'm your assistant, what can I do for you?"
+                prompt.opener = "Hi! I'm your assistant. What can I do for you?"
             if prompt.prompt is None:
                 prompt.prompt = (
                     "You are an intelligent assistant. Please summarize the content of the knowledge base to answer the question. "

sdk/python/test/test_http_api/test_chat_assistant_management/test_create_chat_assistant.py

@@ -221,7 +221,7 @@ class TestChatAssistantCreate:
                 assert res["data"]["prompt"]["variables"] == [{"key": "knowledge", "optional": False}]
                 assert res["data"]["prompt"]["rerank_model"] == ""
                 assert res["data"]["prompt"]["empty_response"] == "Sorry! No relevant content was found in the knowledge base!"
-                assert res["data"]["prompt"]["opener"] == "Hi! I'm your assistant, what can I do for you?"
+                assert res["data"]["prompt"]["opener"] == "Hi! I'm your assistant. What can I do for you?"
                 assert res["data"]["prompt"]["show_quote"] is True
                 assert (
                     res["data"]["prompt"]["prompt"]

sdk/python/test/test_http_api/test_chat_assistant_management/test_update_chat_assistant.py

@@ -218,7 +218,7 @@ class TestChatAssistantUpdate:
                 assert res["data"]["prompt"][0]["variables"] == [{"key": "knowledge", "optional": False}]
                 assert res["data"]["prompt"][0]["rerank_model"] == ""
                 assert res["data"]["prompt"][0]["empty_response"] == "Sorry! No relevant content was found in the knowledge base!"
-                assert res["data"]["prompt"][0]["opener"] == "Hi! I'm your assistant, what can I do for you?"
+                assert res["data"]["prompt"][0]["opener"] == "Hi! I'm your assistant. What can I do for you?"
                 assert res["data"]["prompt"][0]["show_quote"] is True
                 assert (
                     res["data"]["prompt"][0]["prompt"]

test/testcases/test_http_api/test_chat_assistant_management/test_create_chat_assistant.py

@@ -222,7 +222,7 @@ class TestChatAssistantCreate:
                 assert res["data"]["prompt"]["variables"] == [{"key": "knowledge", "optional": False}]
                 assert res["data"]["prompt"]["rerank_model"] == ""
                 assert res["data"]["prompt"]["empty_response"] == "Sorry! No relevant content was found in the knowledge base!"
-                assert res["data"]["prompt"]["opener"] == "Hi! I'm your assistant, what can I do for you?"
+                assert res["data"]["prompt"]["opener"] == "Hi! I'm your assistant. What can I do for you?"
                 assert res["data"]["prompt"]["show_quote"] is True
                 assert (
                     res["data"]["prompt"]["prompt"]

test/testcases/test_http_api/test_chat_assistant_management/test_update_chat_assistant.py

@@ -219,7 +219,7 @@ class TestChatAssistantUpdate:
                 assert res["data"]["prompt"][0]["variables"] == [{"key": "knowledge", "optional": False}]
                 assert res["data"]["prompt"][0]["rerank_model"] == ""
                 assert res["data"]["prompt"][0]["empty_response"] == "Sorry! No relevant content was found in the knowledge base!"
-                assert res["data"]["prompt"][0]["opener"] == "Hi! I'm your assistant, what can I do for you?"
+                assert res["data"]["prompt"][0]["opener"] == "Hi! I'm your assistant. What can I do for you?"
                 assert res["data"]["prompt"][0]["show_quote"] is True
                 assert (
                     res["data"]["prompt"][0]["prompt"]

test/testcases/test_http_api/test_chunk_management_within_dataset/test_update_chunk.py

@@ -245,4 +245,4 @@ class TestUpdatedChunk:
         delete_documents(HttpApiAuth, dataset_id, {"ids": [document_id]})
         res = update_chunk(HttpApiAuth, dataset_id, document_id, chunk_ids[0])
         assert res["code"] == 102
-        assert res["message"] == f"Can't find this chunk {chunk_ids[0]}"
+        assert res["message"] in [f"You don't own the document {document_id}.", f"Can't find this chunk {chunk_ids[0]}"]

test/testcases/test_http_api/test_dataset_mangement/test_list_datasets.py

@@ -163,9 +163,9 @@ class TestDatasetsList:
         [
             {"orderby": ""},
             {"orderby": "unknown"},
-            ({"orderby": "CREATE_TIME"}, lambda r: (is_sorted(r["data"], "create_time", True))),
-            ({"orderby": "UPDATE_TIME"}, lambda r: (is_sorted(r["data"], "update_time", True))),
-            ({"orderby": " create_time "}, lambda r: (is_sorted(r["data"], "update_time", True))),
+            {"orderby": "CREATE_TIME"},
+            {"orderby": "UPDATE_TIME"},
+            {"orderby": " create_time "},
         ],
         ids=["empty", "unknown", "orderby_create_time_upper", "orderby_update_time_upper", "whitespace"],
     )

test/testcases/test_sdk_api/test_chat_assistant_management/test_create_chat_assistant.py

@@ -207,7 +207,7 @@ class TestChatAssistantCreate:
                 assert attrgetter("variables")(chat_assistant.prompt) == [{"key": "knowledge", "optional": False}]
                 assert attrgetter("rerank_model")(chat_assistant.prompt) == ""
                 assert attrgetter("empty_response")(chat_assistant.prompt) == "Sorry! No relevant content was found in the knowledge base!"
-                assert attrgetter("opener")(chat_assistant.prompt) == "Hi! I'm your assistant, what can I do for you?"
+                assert attrgetter("opener")(chat_assistant.prompt) == "Hi! I'm your assistant. What can I do for you?"
                 assert attrgetter("show_quote")(chat_assistant.prompt) is True
                 assert (
                     attrgetter("prompt")(chat_assistant.prompt)

test/testcases/test_sdk_api/test_chat_assistant_management/test_update_chat_assistant.py

@@ -200,7 +200,7 @@ class TestChatAssistantUpdate:
                         "variables": [{"key": "knowledge", "optional": False}],
                         "rerank_model": "",
                         "empty_response": "Sorry! No relevant content was found in the knowledge base!",
-                        "opener": "Hi! I'm your assistant, what can I do for you?",
+                        "opener": "Hi! I'm your assistant. What can I do for you?",
                         "show_quote": True,
                         "prompt": 'You are an intelligent assistant. Please summarize the content of the knowledge base to answer the question. Please list the data in the knowledge base and answer in detail. When all knowledge base content is irrelevant to the question, your answer must include the sentence "The answer you are looking for is not found in the knowledge base!" Answers need to consider chat history.\n      Here is the knowledge base:\n      {knowledge}\n      The above is the knowledge base.',
                     },

test/testcases/test_sdk_api/test_chunk_management_within_dataset/test_update_chunk.py

@@ -151,4 +151,4 @@ class TestUpdatedChunk:
 
         with pytest.raises(Exception) as excinfo:
             chunks[0].update({})
-        assert f"Can't find this chunk {chunks[0].id}" in str(excinfo.value), str(excinfo.value)
+        assert str(excinfo.value) in [f"You don't own the document {chunks[0].document_id}", f"Can't find this chunk {chunks[0].id}"], str(excinfo.value)

test/testcases/test_sdk_api/test_dataset_mangement/test_update_dataset.py

@@ -693,6 +693,7 @@ class TestDatasetUpdate:
             client,
             {
                 "raptor": {"use_raptor": False},
+                "graphrag": {"use_graphrag": False},
             },
         )
         dataset.update({"chunk_method": "qa"})
@@ -708,6 +709,7 @@ class TestDatasetUpdate:
             client,
             {
                 "raptor": {"use_raptor": False},
+                "graphrag": {"use_graphrag": False},
             },
         )
         dataset.update({"chunk_method": "qa", "parser_config": None})

uv.lock

@@ -1,4 +1,5 @@
 version = 1
+revision = 1
 requires-python = ">=3.10, <3.13"
 resolution-markers = [
     "python_full_version >= '3.12' and sys_platform == 'darwin'",
@@ -30,6 +31,15 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/9f/1c/a17fb513aeb684fb83bef5f395910f53103ab30308bbdd77fd66d6698c46/accelerate-1.9.0-py3-none-any.whl", hash = "sha256:c24739a97ade1d54af4549a65f8b6b046adc87e2b3e4d6c66516e32c53d5a8f1" },
 ]
 
+[[package]]
+name = "aiofiles"
+version = "24.1.0"
+source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/0b/03/a88171e277e8caa88a4c77808c20ebb04ba74cc4681bf1e9416c862de237/aiofiles-24.1.0.tar.gz", hash = "sha256:22a075c9e5a3810f0c2e48f3008c94d68c65d763b9b03857924c99e57355166c" }
+wheels = [
+    { url = "https://mirrors.aliyun.com/pypi/packages/a5/45/30bb92d442636f570cb5651bc661f52b610e2eec3f891a5dc3a4c3667db0/aiofiles-24.1.0-py3-none-any.whl", hash = "sha256:b4ec55f4195e3eb5d7abd1bf7e061763e864dd4954231fb8539a0ef8bb8260e5" },
+]
+
 [[package]]
 name = "aiohappyeyeballs"
 version = "2.6.1"
@@ -1028,24 +1038,29 @@ wheels = [
 
 [[package]]
 name = "crawl4ai"
-version = "0.3.8"
+version = "0.3.745"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
+    { name = "aiofiles" },
     { name = "aiosqlite" },
     { name = "beautifulsoup4" },
+    { name = "colorama" },
     { name = "html2text" },
     { name = "litellm" },
     { name = "lxml" },
     { name = "numpy" },
     { name = "pillow" },
     { name = "playwright" },
-    { name = "playwright-stealth" },
     { name = "python-dotenv" },
+    { name = "rank-bm25" },
     { name = "requests" },
+    { name = "snowballstemmer" },
+    { name = "tf-playwright-stealth" },
+    { name = "xxhash" },
 ]
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/1c/31/327598a0c2cc3cd13dcb786ab41e9638c4c100db1940c9345b1e4d953f39/crawl4ai-0.3.8.tar.gz", hash = "sha256:bacc97509ddbfa5e328e299538a27a4c7fc2317e3fd5ad707b04677e4fc23fc6" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/02/5a/919e64ff2977d7aa1b2cda4d45f16ff8996cd2c2dc1f55936fb6cd214222/crawl4ai-0.3.745.tar.gz", hash = "sha256:990396d57e10ae7ccabf35c34a317dbd8c59a3ceca475eac75320a8808334438" }
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/af/03/4d69b8d64b39096a721808a349199ca5d7989acf2177e270d15e6f82c356/Crawl4AI-0.3.8-py3-none-any.whl", hash = "sha256:aa19165440c32b667b7325c166d68b00a99375b09e3a7db929d3873064d5ef4f" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/ed/7e/ebe351a457140330b20b6d8289b8f243b21de6e6bce505cd15b230a83bcb/Crawl4AI-0.3.745-py3-none-any.whl", hash = "sha256:763e6aba80959e60e1fe70cb9d954a4cf257eb230af30f51fcd99ff641a7a88d" },
 ]
 
 [[package]]
@@ -1175,9 +1190,6 @@ name = "datrie"
 version = "0.8.2"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 sdist = { url = "https://mirrors.aliyun.com/pypi/packages/9d/fe/db74bd405d515f06657f11ad529878fd389576dca4812bea6f98d9b31574/datrie-0.8.2.tar.gz", hash = "sha256:525b08f638d5cf6115df6ccd818e5a01298cd230b2dac91c8ff2e6499d18765d" }
-wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/44/02/53f0cf0bf0cd629ba6c2cc13f2f9db24323459e9c19463783d890a540a96/datrie-0.8.2-pp273-pypy_73-win32.whl", hash = "sha256:b07bd5fdfc3399a6dab86d6e35c72b1dbd598e80c97509c7c7518ab8774d3fda" },
-]
 
 [[package]]
 name = "debugpy"
@@ -1423,6 +1435,14 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/36/f4/c6e662dade71f56cd2f3735141b265c3c79293c109549c1e6933b0651ffc/exceptiongroup-1.3.0-py3-none-any.whl", hash = "sha256:4d111e6e0c13d0644cad6ddaa7ed0261a0b36971f6d23e7ec9b4b9097da78a10" },
 ]
 
+[[package]]
+name = "fake-http-header"
+version = "0.3.5"
+source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
+wheels = [
+    { url = "https://mirrors.aliyun.com/pypi/packages/e3/0b/2849c87d9f13766e29c0a2f4d31681aa72e035016b251ab19d99bde7b592/fake_http_header-0.3.5-py3-none-any.whl", hash = "sha256:cd05f4bebf1b7e38b5f5c03d7fb820c0c17e87d9614fbee0afa39c32c7a2ad3c" },
+]
+
 [[package]]
 name = "fake-useragent"
 version = "1.5.1"
@@ -1486,17 +1506,17 @@ name = "fastembed-gpu"
 version = "0.3.6"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
-    { name = "huggingface-hub" },
-    { name = "loguru" },
-    { name = "mmh3" },
-    { name = "numpy" },
-    { name = "onnxruntime-gpu" },
-    { name = "pillow" },
-    { name = "pystemmer" },
-    { name = "requests" },
-    { name = "snowballstemmer" },
-    { name = "tokenizers" },
-    { name = "tqdm" },
+    { name = "huggingface-hub", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "loguru", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "mmh3", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "numpy", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "onnxruntime-gpu", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "pillow", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "pystemmer", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "requests", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "snowballstemmer", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "tokenizers", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "tqdm", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
 ]
 sdist = { url = "https://mirrors.aliyun.com/pypi/packages/da/07/7336c7f3d7ee47f33b407eeb50f5eeb152889de538a52a8f1cc637192816/fastembed_gpu-0.3.6.tar.gz", hash = "sha256:ee2de8918b142adbbf48caaffec0c492f864d73c073eea5a3dcd0e8c1041c50d" }
 wheels = [
@@ -1631,6 +1651,19 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/59/f5/67e9cc5c2036f58115f9fe0f00d203cf6780c3ff8ae0e705e7a9d9e8ff9e/Flask_Login-0.6.3-py3-none-any.whl", hash = "sha256:849b25b82a436bf830a054e74214074af59097171562ab10bfa999e6b78aae5d" },
 ]
 
+[[package]]
+name = "flask-mail"
+version = "0.10.0"
+source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
+dependencies = [
+    { name = "blinker" },
+    { name = "flask" },
+]
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/ba/29/e92dc84c675d1e8d260d5768eb3fb65c70cbd33addecf424187587bee862/flask_mail-0.10.0.tar.gz", hash = "sha256:44083e7b02bbcce792209c06252f8569dd5a325a7aaa76afe7330422bd97881d" }
+wheels = [
+    { url = "https://mirrors.aliyun.com/pypi/packages/e4/c0/a81083da779f482494d49195d8b6c9fde21072558253e4a9fb2ec969c3c1/flask_mail-0.10.0-py3-none-any.whl", hash = "sha256:a451e490931bb3441d9b11ebab6812a16bfa81855792ae1bf9c1e1e22c4e51e7" },
+]
+
 [[package]]
 name = "flask-session"
 version = "0.8.0"
@@ -2142,37 +2175,37 @@ wheels = [
 
 [[package]]
 name = "greenlet"
-version = "3.0.3"
+version = "3.2.4"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/17/14/3bddb1298b9a6786539ac609ba4b7c9c0842e12aa73aaa4d8d73ec8f8185/greenlet-3.0.3.tar.gz", hash = "sha256:43374442353259554ce33599da8b692d5aa96f8976d567d4badf263371fbe491" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/03/b8/704d753a5a45507a7aab61f18db9509302ed3d0a27ac7e0359ec2905b1a6/greenlet-3.2.4.tar.gz", hash = "sha256:0dca0d95ff849f9a364385f36ab49f50065d76964944638be9691e1832e9f86d" }
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/a6/64/bea53c592e3e45799f7c8039a8ee7d6883c518eafef1fcae60beb776070f/greenlet-3.0.3-cp310-cp310-macosx_11_0_universal2.whl", hash = "sha256:9da2bd29ed9e4f15955dd1595ad7bc9320308a3b766ef7f837e23ad4b4aac31a" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/a6/d6/408ad9603339db28ce334021b1403dfcfbcb7501a435d49698408d928de7/greenlet-3.0.3-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d353cadd6083fdb056bb46ed07e4340b0869c305c8ca54ef9da3421acbdf6881" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/6c/90/5b14670653f7363fb3e1665f8da6d64bd4c31d53a796d09ef69f48be7273/greenlet-3.0.3-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:dca1e2f3ca00b84a396bc1bce13dd21f680f035314d2379c4160c98153b2059b" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/ef/17/e8e72cabfb5a906c0d976d7fbcc88310df292beea0f816efbefdaf694284/greenlet-3.0.3-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:3ed7fb269f15dc662787f4119ec300ad0702fa1b19d2135a37c2c4de6fadfd4a" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/1c/2f/64628f6ae48e05f585e0eb3fb7399b52e240ef99f602107b445bf6be23ef/greenlet-3.0.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:dd4f49ae60e10adbc94b45c0b5e6a179acc1736cf7a90160b404076ee283cf83" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/24/35/945d5b10648fec9b20bcc6df8952d20bb3bba76413cd71c1fdbee98f5616/greenlet-3.0.3-cp310-cp310-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:73a411ef564e0e097dbe7e866bb2dda0f027e072b04da387282b02c308807405" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/74/00/27e2da76b926e9b5a2c97d3f4c0baf1b7d8181209d3026c0171f621ae6c0/greenlet-3.0.3-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:7f362975f2d179f9e26928c5b517524e89dd48530a0202570d55ad6ca5d8a56f" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/e1/65/506e0a80931170b0dac1a03d36b7fc299f3fa3576235b916718602fff2c3/greenlet-3.0.3-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:649dde7de1a5eceb258f9cb00bdf50e978c9db1b996964cd80703614c86495eb" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/a6/76/e1ee9f290bb0d46b09704c2fb0e609cae329eb308ad404c0ee6fa1ecb8a5/greenlet-3.0.3-cp310-cp310-win_amd64.whl", hash = "sha256:68834da854554926fbedd38c76e60c4a2e3198c6fbed520b106a8986445caaf9" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/6e/20/68a278a6f93fa36e21cfc3d7599399a8a831225644eb3b6b18755cd3d6fc/greenlet-3.0.3-cp311-cp311-macosx_11_0_universal2.whl", hash = "sha256:b1b5667cced97081bf57b8fa1d6bfca67814b0afd38208d52538316e9422fc61" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/21/b4/90e06e07c78513ab03855768200bdb35c8e764e805b3f14fb488e56f82dc/greenlet-3.0.3-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:52f59dd9c96ad2fc0d5724107444f76eb20aaccb675bf825df6435acb7703559" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/f6/a2/0ed21078039072f9dc738bbf3af12b103a84106b1385ac4723841f846ce7/greenlet-3.0.3-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:afaff6cf5200befd5cec055b07d1c0a5a06c040fe5ad148abcd11ba6ab9b114e" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/42/11/42ad6b1104c357826bbee7d7b9e4f24dbd9fde94899a03efb004aab62963/greenlet-3.0.3-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:fe754d231288e1e64323cfad462fcee8f0288654c10bdf4f603a39ed923bef33" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/bb/6b/384dee7e0121cbd1757bdc1824a5ee28e43d8d4e3f99aa59521f629442fe/greenlet-3.0.3-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:2797aa5aedac23af156bbb5a6aa2cd3427ada2972c828244eb7d1b9255846379" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/c6/1f/12d5a6cc26e8b483c2e7975f9c22e088ac735c0d8dcb8a8f72d31a4e5f04/greenlet-3.0.3-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:b7f009caad047246ed379e1c4dbcb8b020f0a390667ea74d2387be2998f58a22" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/c7/ec/85b647e59e0f137c7792a809156f413e38379cf7f3f2e1353c37f4be4026/greenlet-3.0.3-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:c5e1536de2aad7bf62e27baf79225d0d64360d4168cf2e6becb91baf1ed074f3" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/94/ed/1e5f4bca691a81700e5a88e86d6f0e538acb10188cd2cc17140e523255ef/greenlet-3.0.3-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:894393ce10ceac937e56ec00bb71c4c2f8209ad516e96033e4b3b1de270e200d" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/47/79/26d54d7d700ef65b689fc2665a40846d13e834da0486674a8d4f0f371a47/greenlet-3.0.3-cp311-cp311-win_amd64.whl", hash = "sha256:1ea188d4f49089fc6fb283845ab18a2518d279c7cd9da1065d7a84e991748728" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/a2/2f/461615adc53ba81e99471303b15ac6b2a6daa8d2a0f7f77fd15605e16d5b/greenlet-3.0.3-cp312-cp312-macosx_11_0_universal2.whl", hash = "sha256:70fb482fdf2c707765ab5f0b6655e9cfcf3780d8d87355a063547b41177599be" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/e9/55/2c3cfa3cdbb940cf7321fbcf544f0e9c74898eed43bf678abf416812d132/greenlet-3.0.3-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d4d1ac74f5c0c0524e4a24335350edad7e5f03b9532da7ea4d3c54d527784f2e" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/38/77/efb21ab402651896c74f24a172eb4d7479f9f53898bd5e56b9e20bb24ffd/greenlet-3.0.3-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:149e94a2dd82d19838fe4b2259f1b6b9957d5ba1b25640d2380bea9c5df37676" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/74/3a/92f188ace0190f0066dca3636cf1b09481d0854c46e92ec5e29c7cefe5b1/greenlet-3.0.3-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:15d79dd26056573940fcb8c7413d84118086f2ec1a8acdfa854631084393efcc" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/63/0f/847ed02cdfce10f0e6e3425cd054296bddb11a17ef1b34681fa01a055187/greenlet-3.0.3-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:881b7db1ebff4ba09aaaeae6aa491daeb226c8150fc20e836ad00041bcb11230" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/bd/37/56b0da468a85e7704f3b2bc045015301bdf4be2184a44868c71f6dca6fe2/greenlet-3.0.3-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:fcd2469d6a2cf298f198f0487e0a5b1a47a42ca0fa4dfd1b6862c999f018ebbf" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/7c/68/b5f4084c0a252d7e9c0d95fc1cfc845d08622037adb74e05be3a49831186/greenlet-3.0.3-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:1f672519db1796ca0d8753f9e78ec02355e862d0998193038c7073045899f305" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/a4/fa/31e22345518adcd69d1d6ab5087a12c178aa7f3c51103f6d5d702199d243/greenlet-3.0.3-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:2516a9957eed41dd8f1ec0c604f1cdc86758b587d964668b5b196a9db5bfcde6" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/53/80/3d94d5999b4179d91bcc93745d1b0815b073d61be79dd546b840d17adb18/greenlet-3.0.3-cp312-cp312-win_amd64.whl", hash = "sha256:bba5387a6975598857d86de9eac14210a49d554a77eb8261cc68b7d082f78ce2" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/7d/ed/6bfa4109fcb23a58819600392564fea69cdc6551ffd5e69ccf1d52a40cbc/greenlet-3.2.4-cp310-cp310-macosx_11_0_universal2.whl", hash = "sha256:8c68325b0d0acf8d91dde4e6f930967dd52a5302cd4062932a6b2e7c2969f47c" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/2a/fc/102ec1a2fc015b3a7652abab7acf3541d58c04d3d17a8d3d6a44adae1eb1/greenlet-3.2.4-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:94385f101946790ae13da500603491f04a76b6e4c059dab271b3ce2e283b2590" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/c5/26/80383131d55a4ac0fb08d71660fd77e7660b9db6bdb4e8884f46d9f2cc04/greenlet-3.2.4-cp310-cp310-manylinux2014_ppc64le.manylinux_2_17_ppc64le.whl", hash = "sha256:f10fd42b5ee276335863712fa3da6608e93f70629c631bf77145021600abc23c" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/9f/7c/e7833dbcd8f376f3326bd728c845d31dcde4c84268d3921afcae77d90d08/greenlet-3.2.4-cp310-cp310-manylinux2014_s390x.manylinux_2_17_s390x.whl", hash = "sha256:c8c9e331e58180d0d83c5b7999255721b725913ff6bc6cf39fa2a45841a4fd4b" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/e9/49/547b93b7c0428ede7b3f309bc965986874759f7d89e4e04aeddbc9699acb/greenlet-3.2.4-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:58b97143c9cc7b86fc458f215bd0932f1757ce649e05b640fea2e79b54cedb31" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/7f/91/ae2eb6b7979e2f9b035a9f612cf70f1bf54aad4e1d125129bef1eae96f19/greenlet-3.2.4-cp310-cp310-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:c2ca18a03a8cfb5b25bc1cbe20f3d9a4c80d8c3b13ba3df49ac3961af0b1018d" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/f7/85/433de0c9c0252b22b16d413c9407e6cb3b41df7389afc366ca204dbc1393/greenlet-3.2.4-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:9fe0a28a7b952a21e2c062cd5756d34354117796c6d9215a87f55e38d15402c5" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/a1/8d/88f3ebd2bc96bf7747093696f4335a0a8a4c5acfcf1b757717c0d2474ba3/greenlet-3.2.4-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:8854167e06950ca75b898b104b63cc646573aa5fef1353d4508ecdd1ee76254f" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/d6/6f/b60b0291d9623c496638c582297ead61f43c4b72eef5e9c926ef4565ec13/greenlet-3.2.4-cp310-cp310-win_amd64.whl", hash = "sha256:73f49b5368b5359d04e18d15828eecc1806033db5233397748f4ca813ff1056c" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/a4/de/f28ced0a67749cac23fecb02b694f6473f47686dff6afaa211d186e2ef9c/greenlet-3.2.4-cp311-cp311-macosx_11_0_universal2.whl", hash = "sha256:96378df1de302bc38e99c3a9aa311967b7dc80ced1dcc6f171e99842987882a2" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/09/16/2c3792cba130000bf2a31c5272999113f4764fd9d874fb257ff588ac779a/greenlet-3.2.4-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:1ee8fae0519a337f2329cb78bd7a8e128ec0f881073d43f023c7b8d4831d5246" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/ae/8f/95d48d7e3d433e6dae5b1682e4292242a53f22df82e6d3dda81b1701a960/greenlet-3.2.4-cp311-cp311-manylinux2014_ppc64le.manylinux_2_17_ppc64le.whl", hash = "sha256:94abf90142c2a18151632371140b3dba4dee031633fe614cb592dbb6c9e17bc3" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/d5/5e/405965351aef8c76b8ef7ad370e5da58d57ef6068df197548b015464001a/greenlet-3.2.4-cp311-cp311-manylinux2014_s390x.manylinux_2_17_s390x.whl", hash = "sha256:4d1378601b85e2e5171b99be8d2dc85f594c79967599328f95c1dc1a40f1c633" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/25/5d/382753b52006ce0218297ec1b628e048c4e64b155379331f25a7316eb749/greenlet-3.2.4-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:0db5594dce18db94f7d1650d7489909b57afde4c580806b8d9203b6e79cdc079" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/1f/8e/abdd3f14d735b2929290a018ecf133c901be4874b858dd1c604b9319f064/greenlet-3.2.4-cp311-cp311-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:2523e5246274f54fdadbce8494458a2ebdcdbc7b802318466ac5606d3cded1f8" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/5d/65/deb2a69c3e5996439b0176f6651e0052542bb6c8f8ec2e3fba97c9768805/greenlet-3.2.4-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:1987de92fec508535687fb807a5cea1560f6196285a4cde35c100b8cd632cc52" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/3f/cc/b07000438a29ac5cfb2194bfc128151d52f333cee74dd7dfe3fb733fc16c/greenlet-3.2.4-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:55e9c5affaa6775e2c6b67659f3a71684de4c549b3dd9afca3bc773533d284fa" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/d8/0f/30aef242fcab550b0b3520b8e3561156857c94288f0332a79928c31a52cf/greenlet-3.2.4-cp311-cp311-win_amd64.whl", hash = "sha256:9c40adce87eaa9ddb593ccb0fa6a07caf34015a29bf8d344811665b573138db9" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/44/69/9b804adb5fd0671f367781560eb5eb586c4d495277c93bde4307b9e28068/greenlet-3.2.4-cp312-cp312-macosx_11_0_universal2.whl", hash = "sha256:3b67ca49f54cede0186854a008109d6ee71f66bd57bb36abd6d0a0267b540cdd" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/46/e9/d2a80c99f19a153eff70bc451ab78615583b8dac0754cfb942223d2c1a0d/greenlet-3.2.4-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:ddf9164e7a5b08e9d22511526865780a576f19ddd00d62f8a665949327fde8bb" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/3b/16/035dcfcc48715ccd345f3a93183267167cdd162ad123cd93067d86f27ce4/greenlet-3.2.4-cp312-cp312-manylinux2014_ppc64le.manylinux_2_17_ppc64le.whl", hash = "sha256:f28588772bb5fb869a8eb331374ec06f24a83a9c25bfa1f38b6993afe9c1e968" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/31/da/0386695eef69ffae1ad726881571dfe28b41970173947e7c558d9998de0f/greenlet-3.2.4-cp312-cp312-manylinux2014_s390x.manylinux_2_17_s390x.whl", hash = "sha256:5c9320971821a7cb77cfab8d956fa8e39cd07ca44b6070db358ceb7f8797c8c9" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/68/88/69bf19fd4dc19981928ceacbc5fd4bb6bc2215d53199e367832e98d1d8fe/greenlet-3.2.4-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:c60a6d84229b271d44b70fb6e5fa23781abb5d742af7b808ae3f6efd7c9c60f6" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/19/0d/6660d55f7373b2ff8152401a83e02084956da23ae58cddbfb0b330978fe9/greenlet-3.2.4-cp312-cp312-manylinux_2_24_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:3b3812d8d0c9579967815af437d96623f45c0f2ae5f04e366de62a12d83a8fb0" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/8e/1a/c953fdedd22d81ee4629afbb38d2f9d71e37d23caace44775a3a969147d4/greenlet-3.2.4-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:abbf57b5a870d30c4675928c37278493044d7c14378350b3aa5d484fa65575f0" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/3f/c7/12381b18e21aef2c6bd3a636da1088b888b97b7a0362fac2e4de92405f97/greenlet-3.2.4-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:20fb936b4652b6e307b8f347665e2c615540d4b42b3b4c8a321d8286da7e520f" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/e9/08/b0814846b79399e585f974bbeebf5580fbe59e258ea7be64d9dfb253c84f/greenlet-3.2.4-cp312-cp312-win_amd64.whl", hash = "sha256:a7d4e128405eea3814a12cc2605e0e6aedb4035bf32697f72deca74de4105e02" },
 ]
 
 [[package]]
@@ -2375,7 +2408,7 @@ wheels = [
 
 [[package]]
 name = "httpx"
-version = "0.27.0"
+version = "0.27.2"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
     { name = "anyio" },
@@ -2384,9 +2417,14 @@ dependencies = [
     { name = "idna" },
     { name = "sniffio" },
 ]
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/5c/2d/3da5bdf4408b8b2800061c339f240c1802f2e82d55e50bd39c5a881f47f0/httpx-0.27.0.tar.gz", hash = "sha256:a0cb88a46f32dc874e04ee956e4c2764aba2aa228f650b06788ba6bda2962ab5" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/78/82/08f8c936781f67d9e6b9eeb8a0c8b4e406136ea4c3d1f89a5db71d42e0e6/httpx-0.27.2.tar.gz", hash = "sha256:f7c2be1d2f3c3c3160d441802406b206c2b76f5947b11115e6df10c6c65e66c2" }
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/41/7b/ddacf6dcebb42466abd03f368782142baa82e08fc0c1f8eaa05b4bae87d5/httpx-0.27.0-py3-none-any.whl", hash = "sha256:71d5465162c13681bff01ad59b2cc68dd838ea1f10e51574bac27103f00c91a5" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/56/95/9377bcb415797e44274b51d46e3249eba641711cf3348050f76ee7b15ffc/httpx-0.27.2-py3-none-any.whl", hash = "sha256:7bb2708e112d8fdd7829cd4243970f0c223274051cb35ee80c03301ee29a3df0" },
+]
+
+[package.optional-dependencies]
+socks = [
+    { name = "socksio" },
 ]
 
 [[package]]
@@ -2857,24 +2895,24 @@ wheels = [
 
 [[package]]
 name = "litellm"
-version = "1.48.0"
+version = "1.75.5.post1"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
     { name = "aiohttp" },
     { name = "click" },
+    { name = "httpx" },
     { name = "importlib-metadata" },
     { name = "jinja2" },
     { name = "jsonschema" },
     { name = "openai" },
     { name = "pydantic" },
     { name = "python-dotenv" },
-    { name = "requests" },
     { name = "tiktoken" },
     { name = "tokenizers" },
 ]
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/85/cf/ec69c348c6f16148a55657f3bd63215e965028441c0f322ae8edf9c1210a/litellm-1.48.0.tar.gz", hash = "sha256:31a9b8a25a9daf44c24ddc08bf74298da920f2c5cea44135e5061278d0aa6fc9" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/10/97/6091a020895102a20f1da204ebe68c1293123555476b38e749f95ba5981c/litellm-1.75.5.post1.tar.gz", hash = "sha256:e40a0e4b25032755dc5df7f02742abe9e3b8836236363f605f3bdd363cb5a0d0" }
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/37/2b/6a42747557dc557e71d1e0664c4d5a814b08cda0589213921bb51c64c5e4/litellm-1.48.0-py3-none-any.whl", hash = "sha256:7765e8a92069778f5fc66aacfabd0e2f8ec8d74fb117f5e475567d89b0d376b9" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/8f/76/780f68a3b26227136a5147c76860aacedcae9bf1b7afc1c991ec9cad11bc/litellm-1.75.5.post1-py3-none-any.whl", hash = "sha256:1c72809a9c8f6e132ad06eb7e628f674c775b0ce6bfb58cbd37e8b585d929cb7" },
 ]
 
 [[package]]
@@ -3765,12 +3803,12 @@ name = "onnxruntime-gpu"
 version = "1.19.2"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
-    { name = "coloredlogs" },
-    { name = "flatbuffers" },
-    { name = "numpy" },
-    { name = "packaging" },
-    { name = "protobuf" },
-    { name = "sympy" },
+    { name = "coloredlogs", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "flatbuffers", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "numpy", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "packaging", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "protobuf", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
+    { name = "sympy", marker = "(platform_machine != 'aarch64' and sys_platform == 'linux') or (sys_platform != 'darwin' and sys_platform != 'linux')" },
 ]
 wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/d0/9c/3fa310e0730643051eb88e884f19813a6c8b67d0fbafcda610d960e589db/onnxruntime_gpu-1.19.2-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:a49740e079e7c5215830d30cde3df792e903df007aa0b0fd7aa797937061b27a" },
@@ -3783,7 +3821,7 @@ wheels = [
 
 [[package]]
 name = "openai"
-version = "1.45.0"
+version = "1.99.9"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
     { name = "anyio" },
@@ -3795,9 +3833,9 @@ dependencies = [
     { name = "tqdm" },
     { name = "typing-extensions" },
 ]
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/70/cd/5ec65b9a56999370c032af7933433143f78239d44a8c03a5ba34159af945/openai-1.45.0.tar.gz", hash = "sha256:731207d10637335413aa3c0955f8f8df30d7636a4a0f9c381f2209d32cf8de97" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/8a/d2/ef89c6f3f36b13b06e271d3cc984ddd2f62508a0972c1cbcc8485a6644ff/openai-1.99.9.tar.gz", hash = "sha256:f2082d155b1ad22e83247c3de3958eb4255b20ccf4a1de2e6681b6957b554e92" }
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/d4/2a/97e80a4551346efc9cd937e11adb640207acc5045fdf4e06786eac55bfb1/openai-1.45.0-py3-none-any.whl", hash = "sha256:2f1f7b7cf90f038a9f1c24f0d26c0f1790c102ec5acd07ffd70a9b7feac1ff4e" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/e8/fb/df274ca10698ee77b07bff952f302ea627cc12dac6b85289485dd77db6de/openai-1.99.9-py3-none-any.whl", hash = "sha256:9dbcdb425553bae1ac5d947147bebbd630d91bbfc7788394d4c4f3a35682ab3a" },
 ]
 
 [[package]]
@@ -4242,32 +4280,21 @@ wheels = [
 
 [[package]]
 name = "playwright"
-version = "1.47.0"
+version = "1.54.0"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
     { name = "greenlet" },
     { name = "pyee" },
 ]
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/f8/70/01cad1d41861cd939fe66bff725771dd03f2de39b7c25b4479de2f583ce0/playwright-1.47.0-py3-none-macosx_10_13_x86_64.whl", hash = "sha256:f205df24edb925db1a4ab62f1ab0da06f14bb69e382efecfb0deedc4c7f4b8cd" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/42/17/2300e578b434b56ebfc3d56a5e0fe6dc5e99d6ff43a88fa492b881f3b7e3/playwright-1.47.0-py3-none-macosx_11_0_arm64.whl", hash = "sha256:7fc820faf6885f69a52ba4ec94124e575d3c4a4003bf29200029b4a4f2b2d0ab" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/5a/6a/3cff2abfa4b4c52e1fa34fa8b71bf09cc2a89b03b7417733e5138f1be61d/playwright-1.47.0-py3-none-macosx_11_0_universal2.whl", hash = "sha256:8e212dc472ff19c7d46ed7e900191c7a786ce697556ac3f1615986ec3aa00341" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/80/a6/c5152c817db664d75c439c2bd99d51f906a31c1df4a04e673ef51008b12f/playwright-1.47.0-py3-none-manylinux1_x86_64.whl", hash = "sha256:a1935672531963e4b2a321de5aa59b982fb92463ee6e1032dd7326378e462955" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/d6/50/b573c13d3748a1ab94ed45f2faeb868c63263df0055f57028c4cc775419f/playwright-1.47.0-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e0a1b61473d6f7f39c5d77d4800b3cbefecb03344c90b98f3fbcae63294ad249" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/7d/6c/34225ee5707db5e34bffa77f05d152c797c0e0b9bf3d3a5b426d99160f8f/playwright-1.47.0-py3-none-win32.whl", hash = "sha256:1b977ed81f6bba5582617684a21adab9bad5676d90a357ebf892db7bdf4a9974" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/cb/88/9a3c77025702e506fe04275e677676246ff0b2e6964de5d2527dfdab3416/playwright-1.47.0-py3-none-win_amd64.whl", hash = "sha256:0ec1056042d2e86088795a503347407570bffa32cbe20748e5d4c93dba085280" },
-]
-
-[[package]]
-name = "playwright-stealth"
-version = "1.0.6"
-source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
-dependencies = [
-    { name = "playwright" },
-]
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/e5/dc/4e88b517e4c9cfb63f1b0b67d59adddcef2dc2fe0883b90e07119d15895a/playwright-stealth-1.0.6.tar.gz", hash = "sha256:b504d951d00fac755c7d13665a29611d415180510bd7d23f14ebc89439ba2043" }
-wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/34/10/60981cb8d8e22487061b98a0803313c4fb519cc95ab1421516304a0cfcd0/playwright_stealth-1.0.6-py3-none-any.whl", hash = "sha256:b1b2bcf58eb6859aa53d42c49b91c4e27b74a6d13fc3d0c85eea513dd55efda3" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/f3/09/33d5bfe393a582d8dac72165a9e88b274143c9df411b65ece1cc13f42988/playwright-1.54.0-py3-none-macosx_10_13_x86_64.whl", hash = "sha256:bf3b845af744370f1bd2286c2a9536f474cc8a88dc995b72ea9a5be714c9a77d" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/e1/7b/51882dc584f7aa59f446f2bb34e33c0e5f015de4e31949e5b7c2c10e54f0/playwright-1.54.0-py3-none-macosx_11_0_arm64.whl", hash = "sha256:780928b3ca2077aea90414b37e54edd0c4bbb57d1aafc42f7aa0b3fd2c2fac02" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/73/a1/7aa8ae175b240c0ec8849fcf000e078f3c693f9aa2ffd992da6550ea0dff/playwright-1.54.0-py3-none-macosx_11_0_universal2.whl", hash = "sha256:81d0b6f28843b27f288cfe438af0a12a4851de57998009a519ea84cee6fbbfb9" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/34/a9/45084fd23b6206f954198296ce39b0acf50debfdf3ec83a593e4d73c9c8a/playwright-1.54.0-py3-none-manylinux1_x86_64.whl", hash = "sha256:09919f45cc74c64afb5432646d7fef0d19fff50990c862cb8d9b0577093f40cc" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/02/d4/6a692f4c6db223adc50a6e53af405b45308db39270957a6afebddaa80ea2/playwright-1.54.0-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:13ae206c55737e8e3eae51fb385d61c0312eeef31535643bb6232741b41b6fdc" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/72/7a/4ee60a1c3714321db187bebbc40d52cea5b41a856925156325058b5fca5a/playwright-1.54.0-py3-none-win32.whl", hash = "sha256:0b108622ffb6906e28566f3f31721cd57dda637d7e41c430287804ac01911f56" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/aa/77/8f8fae05a242ef639de963d7ae70a69d0da61d6d72f1207b8bbf74ffd3e7/playwright-1.54.0-py3-none-win_amd64.whl", hash = "sha256:9e5aee9ae5ab1fdd44cd64153313a2045b136fcbcfb2541cc0a3d909132671a2" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/33/ff/99a6f4292a90504f2927d34032a4baf6adb498dc3f7cf0f3e0e22899e310/playwright-1.54.0-py3-none-win_arm64.whl", hash = "sha256:a975815971f7b8dca505c441a4c56de1aeb56a211290f8cc214eeef5524e8d75" },
 ]
 
 [[package]]
@@ -4653,8 +4680,6 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/59/fe/aae679b64363eb78326c7fdc9d06ec3de18bac68be4b612fc1fe8902693c/pycryptodome-3.23.0-cp37-abi3-win32.whl", hash = "sha256:507dbead45474b62b2bbe318eb1c4c8ee641077532067fec9c1aa82c31f84886" },
     { url = "https://mirrors.aliyun.com/pypi/packages/54/2f/e97a1b8294db0daaa87012c24a7bb714147c7ade7656973fd6c736b484ff/pycryptodome-3.23.0-cp37-abi3-win_amd64.whl", hash = "sha256:c75b52aacc6c0c260f204cbdd834f76edc9fb0d8e0da9fbf8352ef58202564e2" },
     { url = "https://mirrors.aliyun.com/pypi/packages/18/3d/f9441a0d798bf2b1e645adc3265e55706aead1255ccdad3856dbdcffec14/pycryptodome-3.23.0-cp37-abi3-win_arm64.whl", hash = "sha256:11eeeb6917903876f134b56ba11abe95c0b0fd5e3330def218083c7d98bbcb3c" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/9f/7c/f5b0556590e7b4e710509105e668adb55aa9470a9f0e4dea9c40a4a11ce1/pycryptodome-3.23.0-pp27-pypy_73-manylinux2010_x86_64.whl", hash = "sha256:350ebc1eba1da729b35ab7627a833a1a355ee4e852d8ba0447fafe7b14504d56" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/33/38/dcc795578d610ea1aaffef4b148b8cafcfcf4d126b1e58231ddc4e475c70/pycryptodome-3.23.0-pp27-pypy_73-win32.whl", hash = "sha256:93837e379a3e5fd2bb00302a47aee9fdf7940d83595be3915752c74033d17ca7" },
     { url = "https://mirrors.aliyun.com/pypi/packages/d9/12/e33935a0709c07de084d7d58d330ec3f4daf7910a18e77937affdb728452/pycryptodome-3.23.0-pp310-pypy310_pp73-macosx_10_15_x86_64.whl", hash = "sha256:ddb95b49df036ddd264a0ad246d1be5b672000f12d6961ea2c267083a5e19379" },
     { url = "https://mirrors.aliyun.com/pypi/packages/22/0b/aa8f9419f25870889bebf0b26b223c6986652bdf071f000623df11212c90/pycryptodome-3.23.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d8e95564beb8782abfd9e431c974e14563a794a4944c29d6d3b7b5ea042110b4" },
     { url = "https://mirrors.aliyun.com/pypi/packages/d4/5e/63f5cbde2342b7f70a39e591dbe75d9809d6338ce0b07c10406f1a140cdc/pycryptodome-3.23.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:14e15c081e912c4b0d75632acd8382dfce45b258667aa3c67caf7a4d4c13f630" },
@@ -4678,8 +4703,6 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/48/7d/0f2b09490b98cc6a902ac15dda8760c568b9c18cfe70e0ef7a16de64d53a/pycryptodomex-3.20.0-cp35-abi3-musllinux_1_1_x86_64.whl", hash = "sha256:7a7a8f33a1f1fb762ede6cc9cbab8f2a9ba13b196bfaf7bc6f0b39d2ba315a43" },
     { url = "https://mirrors.aliyun.com/pypi/packages/b0/1c/375adb14b71ee1c8d8232904e928b3e7af5bbbca7c04e4bec94fe8e90c3d/pycryptodomex-3.20.0-cp35-abi3-win32.whl", hash = "sha256:c39778fd0548d78917b61f03c1fa8bfda6cfcf98c767decf360945fe6f97461e" },
     { url = "https://mirrors.aliyun.com/pypi/packages/b2/e8/1b92184ab7e5595bf38000587e6f8cf9556ebd1bf0a583619bee2057afbd/pycryptodomex-3.20.0-cp35-abi3-win_amd64.whl", hash = "sha256:2a47bcc478741b71273b917232f521fd5704ab4b25d301669879e7273d3586cc" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/e7/c5/9140bb867141d948c8e242013ec8a8011172233c898dfdba0a2417c3169a/pycryptodomex-3.20.0-pp27-pypy_73-manylinux2010_x86_64.whl", hash = "sha256:1be97461c439a6af4fe1cf8bf6ca5936d3db252737d2f379cc6b2e394e12a458" },
-    { url = "https://mirrors.aliyun.com/pypi/packages/5e/6a/04acb4978ce08ab16890c70611ebc6efd251681341617bbb9e53356dee70/pycryptodomex-3.20.0-pp27-pypy_73-win32.whl", hash = "sha256:19764605feea0df966445d46533729b645033f134baeb3ea26ad518c9fdf212c" },
     { url = "https://mirrors.aliyun.com/pypi/packages/eb/df/3f1ea084e43b91e6d2b6b3493cc948864c17ea5d93ff1261a03812fbfd1a/pycryptodomex-3.20.0-pp310-pypy310_pp73-macosx_10_9_x86_64.whl", hash = "sha256:f2e497413560e03421484189a6b65e33fe800d3bd75590e6d78d4dfdb7accf3b" },
     { url = "https://mirrors.aliyun.com/pypi/packages/c9/f3/83ffbdfa0c8f9154bcd8866895f6cae5a3ec749da8b0840603cf936c4412/pycryptodomex-3.20.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e48217c7901edd95f9f097feaa0388da215ed14ce2ece803d3f300b4e694abea" },
     { url = "https://mirrors.aliyun.com/pypi/packages/c9/9d/c113e640aaf02af5631ae2686b742aac5cd0e1402b9d6512b1c7ec5ef05d/pycryptodomex-3.20.0-pp310-pypy310_pp73-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d00fe8596e1cc46b44bf3907354e9377aa030ec4cd04afbbf6e899fc1e2a7781" },
@@ -4792,14 +4815,14 @@ wheels = [
 
 [[package]]
 name = "pyee"
-version = "12.0.0"
+version = "13.0.0"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
     { name = "typing-extensions" },
 ]
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/d2/a7/8faaa62a488a2a1e0d56969757f087cbd2729e9bcfa508c230299f366b4c/pyee-12.0.0.tar.gz", hash = "sha256:c480603f4aa2927d4766eb41fa82793fe60a82cbfdb8d688e0d08c55a534e145" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/95/03/1fd98d5841cd7964a27d729ccf2199602fe05eb7a405c1462eb7277945ed/pyee-13.0.0.tar.gz", hash = "sha256:b391e3c5a434d1f5118a25615001dbc8f669cf410ab67d04c4d4e07c55481c37" }
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/1d/0d/95993c08c721ec68892547f2117e8f9dfbcef2ca71e098533541b4a54d5f/pyee-12.0.0-py3-none-any.whl", hash = "sha256:7b14b74320600049ccc7d0e0b1becd3b4bd0a03c745758225e31a59f4095c990" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/9b/4d/b9add7c84060d4c1906abe9a7e5359f2a60f7a9a4f67268b2766673427d8/pyee-13.0.0-py3-none-any.whl", hash = "sha256:48195a3cddb3b1515ce0695ed76036b5ccc2ef3a9f963ff9f77aec0139845498" },
 ]
 
 [[package]]
@@ -4915,14 +4938,14 @@ wheels = [
 
 [[package]]
 name = "pypdf"
-version = "5.9.0"
+version = "6.0.0"
 source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
 dependencies = [
     { name = "typing-extensions", marker = "python_full_version < '3.11'" },
 ]
-sdist = { url = "https://mirrors.aliyun.com/pypi/packages/89/3a/584b97a228950ed85aec97c811c68473d9b8d149e6a8c155668287cf1a28/pypdf-5.9.0.tar.gz", hash = "sha256:30f67a614d558e495e1fbb157ba58c1de91ffc1718f5e0dfeb82a029233890a1" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/20/ac/a300a03c3b34967c050677ccb16e7a4b65607ee5df9d51e8b6d713de4098/pypdf-6.0.0.tar.gz", hash = "sha256:282a99d2cc94a84a3a3159f0d9358c0af53f85b4d28d76ea38b96e9e5ac2a08d" }
 wheels = [
-    { url = "https://mirrors.aliyun.com/pypi/packages/48/d9/6cff57c80a6963e7dd183bf09e9f21604a77716644b1e580e97b259f7612/pypdf-5.9.0-py3-none-any.whl", hash = "sha256:be10a4c54202f46d9daceaa8788be07aa8cd5ea8c25c529c50dd509206382c35" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/2c/83/2cacc506eb322bb31b747bc06ccb82cc9aa03e19ee9c1245e538e49d52be/pypdf-6.0.0-py3-none-any.whl", hash = "sha256:56ea60100ce9f11fc3eec4f359da15e9aec3821b036c1f06d2b660d35683abb8" },
 ]
 
 [[package]]
@@ -5026,6 +5049,63 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/30/3d/64ad57c803f1fa1e963a7946b6e0fea4a70df53c1a7fed304586539c2bac/pytest-8.3.5-py3-none-any.whl", hash = "sha256:c69214aa47deac29fad6c2a4f590b9c4a9fdb16a403176fe154b79c0b4d4d820" },
 ]
 
+[[package]]
+name = "python-calamine"
+version = "0.4.0"
+source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
+dependencies = [
+    { name = "packaging" },
+]
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/cc/03/269f96535705b2f18c8977fa58e76763b4e4727a9b3ae277a9468c8ffe05/python_calamine-0.4.0.tar.gz", hash = "sha256:94afcbae3fec36d2d7475095a59d4dc6fae45829968c743cb799ebae269d7bbf" }
+wheels = [
+    { url = "https://mirrors.aliyun.com/pypi/packages/8e/06/885c73fd472cb76af4c4650174a7c11b77a8bc40585044bc445ac694e5e6/python_calamine-0.4.0-cp310-cp310-macosx_10_12_x86_64.whl", hash = "sha256:06011f11fd8d2dbfe0bc9bd8bd135c191aafe66f2d0c9eecf0ae3cb38f42f888" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/2d/76/28c704d875046cc1ca92d8c6e680f6bc38d83735397fee821929691fd57f/python_calamine-0.4.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:12e350e5967bf3206a8b472d9b6c348ff37ae791dba1a1715e076b2c39328557" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/ff/71/dc00e6d9187044d15c85cd0875d8aba2b3e0d3051ceabd47d859f40f69c8/python_calamine-0.4.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:35be298f69006e86b0311a538c1c9694ce3012237c33572d3dfe2bea6b5b9820" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/4b/f6/be5e35263ceec21e77810d7900235124a9a83fd3c0afbbbb79da658d535c/python_calamine-0.4.0-cp310-cp310-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:7abb10367aea435ca473b9b698636db912f2ab164f19a6c9675710ed926f33ac" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/9c/61/0068297ec0000b2c0755a608c8068a1070b8193675d2ff603d390291aa45/python_calamine-0.4.0-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:58c2c4440982ec6db64c826136661f84f84bc0d8ee0cdd64a38128cd217797eb" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/d1/1d/8a9c7d491d31db1def335f4d90b85ea29940d84c59a27a88a442b840fda7/python_calamine-0.4.0-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e58cd89154fd1b5ef77c609f63dce108d390ece5a5f3225ca3ebedc8d343e9d5" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/40/87/a104c11b320d3fb7f1997d97207addce0fe5e1d41e5fd2e8adb0fb8b1325/python_calamine-0.4.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1f90f85e04c281d96c6dc5551176fc4e32c95257c3a2d384a947b3e68275c7d6" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/8c/a5/9580de7758950b39c5048787909b639771c92ad6ea7f909a48dca1dbc6fe/python_calamine-0.4.0-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:d5f8408b01d8097b2e662d0205ca09695788fb5f3492ade27de4ad4160cb6bd4" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/25/a1/2b8282ec4acdf1879f9d46d84a6907843d2a331639719a2cfc90356345b5/python_calamine-0.4.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:3d61957c10d37e6bf508fafdf52e6bb3112db8196e30bca8bc4b4560db2cc5f5" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/95/03/ae56667a2a2eb273eea5f754212454c7073f8abe8e0fdca0edfbe5c0cf37/python_calamine-0.4.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:a5a58bbfcad9c1192dada189e367ed46e72037fcaec585e970fa919b92e07a57" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/52/60/d7aaa39977ab401de33ae975dd704805ec9c76117f0fe3a53e45b718822c/python_calamine-0.4.0-cp310-cp310-win32.whl", hash = "sha256:f06415096bcd9218b6c15d39ee2006ec0f32282e3d08605391d2a8a52187f9ca" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/6a/f5/fdfeccd7d66e5c9c834288df6a657858a046d94a6e4cd624418cb5bb96dd/python_calamine-0.4.0-cp310-cp310-win_amd64.whl", hash = "sha256:e457d1e07acb2798b72e70bd4e88f07cd486ca5129a19fadc6aa19a2cd4e76e8" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/d4/a5/bcd82326d0ff1ab5889e7a5e13c868b483fc56398e143aae8e93149ba43b/python_calamine-0.4.0-cp311-cp311-macosx_10_12_x86_64.whl", hash = "sha256:d1687f8c4d7852920c7b4e398072f183f88dd273baf5153391edc88b7454b8c0" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/f6/1a/a681f1d2f28164552e91ef47bcde6708098aa64a5f5fe3952f22362d340a/python_calamine-0.4.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:258d04230bebbbafa370a15838049d912d6a0a2c4da128943d8160ca4b6db58e" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/3d/92/2fc911431733739d4e7a633cefa903fa49a6b7a61e8765bad29a4a7c47b1/python_calamine-0.4.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:c686e491634934f059553d55f77ac67ca4c235452d5b444f98fe79b3579f1ea5" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/f4/f0/48bfae6802eb360028ca6c15e9edf42243aadd0006b6ac3e9edb41a57119/python_calamine-0.4.0-cp311-cp311-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:4480af7babcc2f919c638a554b06b7b145d9ab3da47fd696d68c2fc6f67f9541" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/a4/dc/f8c956e15bac9d5d1e05cd1b907ae780e40522d2fd103c8c6e2f21dff4ed/python_calamine-0.4.0-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:e405b87a8cd1e90a994e570705898634f105442029f25bab7da658ee9cbaa771" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/54/3f/e69ab97c7734fb850fba2f506b775912fd59f04e17488582c8fbf52dbc72/python_calamine-0.4.0-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:a831345ee42615f0dfcb0ed60a3b1601d2f946d4166edae64fd9a6f9bbd57fc1" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/79/03/b4c056b468908d87a3de94389166e0f4dba725a70bc39e03bc039ba96f6b/python_calamine-0.4.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9951b8e4cafb3e1623bb5dfc31a18d38ef43589275f9657e99dfcbe4c8c4b33e" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/86/4f/b9092f7c970894054083656953184e44cb2dadff8852425e950d4ca419af/python_calamine-0.4.0-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:a6619fe3b5c9633ed8b178684605f8076c9d8d85b29ade15f7a7713fcfdee2d0" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/64/da/137239027bf253aabe7063450950085ec9abd827d0cbc5170f585f38f464/python_calamine-0.4.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:2cc45b8e76ee331f6ea88ca23677be0b7a05b502cd4423ba2c2bc8dad53af1be" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/80/96/74c38bcf6b6825d5180c0e147b85be8c52dbfba11848b1e98ba358e32a64/python_calamine-0.4.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:1b2cfb7ced1a7c80befa0cfddfe4aae65663eb4d63c4ae484b9b7a80ebe1b528" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/33/95/9d7b8fe8b32d99a6c79534df3132cfe40e9df4a0f5204048bf5e66ddbd93/python_calamine-0.4.0-cp311-cp311-win32.whl", hash = "sha256:04f4e32ee16814fc1fafc49300be8eeb280d94878461634768b51497e1444bd6" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/7c/e3/1c6cd9fd499083bea6ff1c30033ee8215b9f64e862babf5be170cacae190/python_calamine-0.4.0-cp311-cp311-win_amd64.whl", hash = "sha256:a8543f69afac2213c0257bb56215b03dadd11763064a9d6b19786f27d1bef586" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/94/1c/3105d19fbab6b66874ce8831652caedd73b23b72e88ce18addf8ceca8c12/python_calamine-0.4.0-cp311-cp311-win_arm64.whl", hash = "sha256:54622e35ec7c3b6f07d119da49aa821731c185e951918f152c2dbf3bec1e15d6" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/63/60/f951513aaaa470b3a38a87d65eca45e0a02bc329b47864f5a17db563f746/python_calamine-0.4.0-cp312-cp312-macosx_10_12_x86_64.whl", hash = "sha256:74bca5d44a73acf3dcfa5370820797fcfd225c8c71abcddea987c5b4f5077e98" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/76/3f/789955bbc77831c639890758f945eb2b25d6358065edf00da6751226cf31/python_calamine-0.4.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:cf80178f5d1b0ee2ccfffb8549c50855f6249e930664adc5807f4d0d6c2b269c" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/00/4c/f87d17d996f647030a40bfd124fe45fe893c002bee35ae6aca9910a923ae/python_calamine-0.4.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:65cfef345386ae86f7720f1be93495a40fd7e7feabb8caa1df5025d7fbc58a1f" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/47/d2/3269367303f6c0488cf1bfebded3f9fe968d118a988222e04c9b2636bf2e/python_calamine-0.4.0-cp312-cp312-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:f23e6214dbf9b29065a5dcfd6a6c674dd0e251407298c9138611c907d53423ff" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/f9/6d/c7ac35f5c7125e8bd07eb36773f300fda20dd2da635eae78a8cebb0b6ab7/python_calamine-0.4.0-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:d792d304ee232ab01598e1d3ab22e074a32c2511476b5fb4f16f4222d9c2a265" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/f0/81/5ea8792a2e9ab5e2a05872db3a4d3ed3538ad5af1861282c789e2f13a8cf/python_calamine-0.4.0-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:bf813425918fd68f3e991ef7c4b5015be0a1a95fc4a8ab7e73c016ef1b881bb4" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/cc/6e/989e56e6f073fc0981a74ba7a393881eb351bb143e5486aa629b5e5d6a8b/python_calamine-0.4.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:bbe2a0ccb4d003635888eea83a995ff56b0748c8c76fc71923544f5a4a7d4cd7" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/5d/92/2c9bd64277c6fe4be695d7d5a803b38d953ec8565037486be7506642c27c/python_calamine-0.4.0-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:a7b3bb5f0d910b9b03c240987560f843256626fd443279759df4e91b717826d2" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/64/fa/fc758ca37701d354a6bc7d63118699f1c73788a1f2e1b44d720824992764/python_calamine-0.4.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:bd2c0fc2b5eabd08ceac8a2935bffa88dbc6116db971aa8c3f244bad3fd0f644" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/65/52/40d7e08ae0ddba331cdc9f7fb3e92972f8f38d7afbd00228158ff6d1fceb/python_calamine-0.4.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:85b547cb1c5b692a0c2406678d666dbc1cec65a714046104683fe4f504a1721d" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/16/de/e8a071c0adfda73285d891898a24f6e99338328c404f497ff5b0e6bc3d45/python_calamine-0.4.0-cp312-cp312-win32.whl", hash = "sha256:4c2a1e3a0db4d6de4587999a21cc35845648c84fba81c03dd6f3072c690888e4" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/5e/f2/7fdfada13f80db12356853cf08697ff4e38800a1809c2bdd26ee60962e7a/python_calamine-0.4.0-cp312-cp312-win_amd64.whl", hash = "sha256:b193c89ffcc146019475cd121c552b23348411e19c04dedf5c766a20db64399a" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/20/66/d37412ad854480ce32f50d9f74f2a2f88b1b8a6fbc32f70aabf3211ae89e/python_calamine-0.4.0-cp312-cp312-win_arm64.whl", hash = "sha256:43a0f15e0b60c75a71b21a012b911d5d6f5fa052afad2a8edbc728af43af0fcf" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/a1/8c/120a1128ff422dba43d6f2d1d19520bca95abf1e5135bbe3b84a782d3927/python_calamine-0.4.0-pp310-pypy310_pp73-macosx_10_12_x86_64.whl", hash = "sha256:4f9a44015de9e19a876babf707dc55708881930024220c8ae926ea0255f705fe" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/db/62/6062945b8d5fc73d0a0c44b25ee5f4037cc32d62b57688a0d0ca6763006d/python_calamine-0.4.0-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:2c10bb42e0d0810368e78ee9359902f999a1f09bcc2391b060f91f981f75ae21" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/9f/7b/597470e5349056a1dd7b0e3a7e838da53cba9186771ca5501a264446f4ad/python_calamine-0.4.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:494c5dd1dcee25935ff9d7a9eef6b0f629266d1670aef3ee5e0d38370dcb3352" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/4b/e9/fabdd376713409e6ae6a8fee41293304798d794a64c9d407ba90f765f3d4/python_calamine-0.4.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5f04fb24e70ab4403fc367b9b779eaa3bf61c140908d9115ddfe1e221372d5d4" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/80/8e/cc09cc14276662cea1f161a20bdce46d8bfd409e84027a0a0515a00b63f5/python_calamine-0.4.0-pp310-pypy310_pp73-manylinux_2_5_i686.manylinux1_i686.whl", hash = "sha256:49dce56dbd1efc024b63b913595f1a9bef6f66a6467aefad7dcd548654fedb5b" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/7b/b4/44f560b0ab4dcb49845f5c5361641885f8dc2b7e9b35fbf59242191c2eeb/python_calamine-0.4.0-pp310-pypy310_pp73-musllinux_1_1_aarch64.whl", hash = "sha256:7dc1755cd0b10ce5e2d80e77e9f19c13ed405b354178c3547ba5a11d34fce6ea" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/9b/4b/ee4ac45d500bd2ae189f84adf3338dbd32579fa2198a05ad180666578575/python_calamine-0.4.0-pp310-pypy310_pp73-musllinux_1_1_x86_64.whl", hash = "sha256:9d981efa53ddfd8733555ad4c9368c8c1254bd8d1e162c93b8d341ead6acc5a9" },
+    { url = "https://mirrors.aliyun.com/pypi/packages/74/99/835a9cd3ed503cb51fd15039b6404c536d201a6f3d16a6e069ce1079c5e4/python_calamine-0.4.0-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:b370998567de0cd7a36a8ac73acabefea8397ad2d9aad3cf245b5d35f74cb990" },
+]
+
 [[package]]
 name = "python-dateutil"
 version = "2.8.2"
@@ -5225,6 +5305,7 @@ dependencies = [
     { name = "flask" },
     { name = "flask-cors" },
     { name = "flask-login" },
+    { name = "flask-mail" },
     { name = "flask-session" },
     { name = "google-generativeai" },
     { name = "google-search-results" },
@@ -5232,13 +5313,14 @@ dependencies = [
     { name = "groq" },
     { name = "hanziconv" },
     { name = "html-text" },
-    { name = "httpx" },
+    { name = "httpx", extra = ["socks"] },
     { name = "huggingface-hub" },
     { name = "infinity-emb" },
     { name = "infinity-sdk" },
     { name = "itsdangerous" },
     { name = "json-repair" },
     { name = "langfuse" },
+    { name = "litellm" },
     { name = "markdown" },
     { name = "markdown-to-json" },
     { name = "mcp" },
@@ -5271,6 +5353,7 @@ dependencies = [
     { name = "pyodbc" },
     { name = "pypdf" },
     { name = "pypdf2" },
+    { name = "python-calamine" },
     { name = "python-dateutil" },
     { name = "python-docx" },
     { name = "python-dotenv" },
@@ -5357,7 +5440,7 @@ requires-dist = [
     { name = "click", specifier = ">=8.1.8" },
     { name = "cn2an", specifier = "==0.5.22" },
     { name = "cohere", specifier = "==5.6.2" },
-    { name = "crawl4ai", specifier = "==0.3.8" },
+    { name = "crawl4ai", specifier = ">=0.3.8" },
     { name = "dashscope", specifier = "==1.20.11" },
     { name = "datrie", specifier = "==0.8.2" },
     { name = "debugpy", specifier = ">=1.8.13" },
@@ -5377,6 +5460,7 @@ requires-dist = [
     { name = "flask", specifier = "==3.0.3" },
     { name = "flask-cors", specifier = "==5.0.0" },
     { name = "flask-login", specifier = "==0.6.3" },
+    { name = "flask-mail", specifier = ">=0.10.0" },
     { name = "flask-session", specifier = "==0.8.0" },
     { name = "google-generativeai", specifier = ">=0.8.1,<0.9.0" },
     { name = "google-search-results", specifier = "==2.4.2" },
@@ -5384,13 +5468,14 @@ requires-dist = [
     { name = "groq", specifier = "==0.9.0" },
     { name = "hanziconv", specifier = "==0.3.2" },
     { name = "html-text", specifier = "==0.6.2" },
-    { name = "httpx", specifier = "==0.27.0" },
+    { name = "httpx", extras = ["socks"], specifier = "==0.27.2" },
     { name = "huggingface-hub", specifier = ">=0.25.0,<0.26.0" },
     { name = "infinity-emb", specifier = ">=0.0.66,<0.0.67" },
     { name = "infinity-sdk", specifier = "==0.6.0.dev4" },
     { name = "itsdangerous", specifier = "==2.1.2" },
     { name = "json-repair", specifier = "==0.35.0" },
     { name = "langfuse", specifier = ">=2.60.0" },
+    { name = "litellm", specifier = ">=1.74.15.post1" },
     { name = "markdown", specifier = "==3.6" },
     { name = "markdown-to-json", specifier = "==2.1.1" },
     { name = "mcp", specifier = ">=1.9.4" },
@@ -5402,7 +5487,7 @@ requires-dist = [
     { name = "ollama", specifier = "==0.2.1" },
     { name = "onnxruntime", marker = "platform_machine != 'x86_64' or sys_platform == 'darwin'", specifier = "==1.19.2" },
     { name = "onnxruntime-gpu", marker = "platform_machine == 'x86_64' and sys_platform != 'darwin'", specifier = "==1.19.2" },
-    { name = "openai", specifier = "==1.45.0" },
+    { name = "openai", specifier = ">=1.45.0" },
     { name = "opencv-python", specifier = "==4.10.0.84" },
     { name = "opencv-python-headless", specifier = "==4.10.0.84" },
     { name = "opendal", specifier = ">=0.45.0,<0.46.0" },
@@ -5421,8 +5506,9 @@ requires-dist = [
     { name = "pyicu", specifier = ">=2.13.1,<3.0.0" },
     { name = "pymysql", specifier = ">=1.1.1,<2.0.0" },
     { name = "pyodbc", specifier = ">=5.2.0,<6.0.0" },
-    { name = "pypdf", specifier = ">=5.0.0,<6.0.0" },
+    { name = "pypdf", specifier = "==6.0.0" },
     { name = "pypdf2", specifier = ">=3.0.1,<4.0.0" },
+    { name = "python-calamine", specifier = ">=0.4.0" },
     { name = "python-dateutil", specifier = "==2.8.2" },
     { name = "python-docx", specifier = ">=1.1.2,<2.0.0" },
     { name = "python-dotenv", specifier = "==1.0.1" },
@@ -5467,6 +5553,7 @@ requires-dist = [
     { name = "yfinance", specifier = "==0.2.65" },
     { name = "zhipuai", specifier = "==2.0.1" },
 ]
+provides-extras = ["full"]
 
 [package.metadata.requires-dev]
 test = [
@@ -5481,6 +5568,18 @@ test = [
     { name = "requests-toolbelt", specifier = ">=1.0.0" },
 ]
 
+[[package]]
+name = "rank-bm25"
+version = "0.2.2"
+source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
+dependencies = [
+    { name = "numpy" },
+]
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/fc/0a/f9579384aa017d8b4c15613f86954b92a95a93d641cc849182467cf0bb3b/rank_bm25-0.2.2.tar.gz", hash = "sha256:096ccef76f8188563419aaf384a02f0ea459503fdf77901378d4fd9d87e5e51d" }
+wheels = [
+    { url = "https://mirrors.aliyun.com/pypi/packages/2a/21/f691fb2613100a62b3fa91e9988c991e9ca5b89ea31c0d3152a3210344f9/rank_bm25-0.2.2-py3-none-any.whl", hash = "sha256:7bd4a95571adadfc271746fa146a4bcfd89c0cf731e49c3d1ad863290adbe8ae" },
+]
+
 [[package]]
 name = "ranx"
 version = "0.3.20"
@@ -6122,6 +6221,15 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/ed/dc/c02e01294f7265e63a7315fe086dd1df7dacb9f840a804da846b96d01b96/snowballstemmer-2.2.0-py2.py3-none-any.whl", hash = "sha256:c8e1716e83cc398ae16824e5572ae04e0d9fc2c6b985fb0f900f5f0c96ecba1a" },
 ]
 
+[[package]]
+name = "socksio"
+version = "1.0.0"
+source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/f8/5c/48a7d9495be3d1c651198fd99dbb6ce190e2274d0f28b9051307bdec6b85/socksio-1.0.0.tar.gz", hash = "sha256:f88beb3da5b5c38b9890469de67d0cb0f9d494b78b106ca1845f96c10b91c4ac" }
+wheels = [
+    { url = "https://mirrors.aliyun.com/pypi/packages/37/c3/6eeb6034408dac0fa653d126c9204ade96b819c936e136c5e8a6897eee9c/socksio-1.0.0-py3-none-any.whl", hash = "sha256:95dc1f15f9b34e8d7b16f06d74b8ccf48f609af32ab33c608d08761c5dcbb1f3" },
+]
+
 [[package]]
 name = "sortedcontainers"
 version = "2.4.0"
@@ -6423,6 +6531,19 @@ wheels = [
     { url = "https://mirrors.aliyun.com/pypi/packages/55/08/98090d1a139e8995053ed22e099b43aa4dea8cffe056f8f0bc5178aeecbd/tencentcloud_sdk_python-3.0.1215-py2.py3-none-any.whl", hash = "sha256:899ced749baf74846f1eabf452f74aa0e48d1965f0ca7828a8b73b446f76f5f2" },
 ]
 
+[[package]]
+name = "tf-playwright-stealth"
+version = "1.2.0"
+source = { registry = "https://mirrors.aliyun.com/pypi/simple" }
+dependencies = [
+    { name = "fake-http-header" },
+    { name = "playwright" },
+]
+sdist = { url = "https://mirrors.aliyun.com/pypi/packages/d6/6b/32bb58c65991f91aeaaf7473b650175d9d4af5dd383983d177d49ccba08d/tf_playwright_stealth-1.2.0.tar.gz", hash = "sha256:7bb8d32d3e60324fbf6b9eeae540b8cd9f3b9e07baeb33b025dbc98ad47658ba" }
+wheels = [
+    { url = "https://mirrors.aliyun.com/pypi/packages/11/3d/2653f4cf49660bb44eeac8270617cc4c0287d61716f249f55053f0af0724/tf_playwright_stealth-1.2.0-py3-none-any.whl", hash = "sha256:26ee47ee89fa0f43c606fe37c188ea3ccd36f96ea90c01d167b768df457e7886" },
+]
+
 [[package]]
 name = "threadpoolctl"
 version = "3.6.0"

web/.eslintrc.js

@@ -1,5 +1,43 @@
 // .eslintrc.js
 module.exports = {
-  // Umi 项目
   extends: [require.resolve('umi/eslint'), 'plugin:react-hooks/recommended'],
+  plugins: ['check-file'],
+  rules: {
+    '@typescript-eslint/no-use-before-define': [
+      'warn',
+      {
+        functions: false,
+        variables: true,
+      },
+    ],
+    'check-file/filename-naming-convention': [
+      'error',
+      {
+        '**/*.{jsx,tsx}': 'KEBAB_CASE',
+        '**/*.{js,ts}': 'KEBAB_CASE',
+      },
+    ],
+    'check-file/folder-naming-convention': [
+      'error',
+      {
+        'src/**/': 'KEBAB_CASE',
+        'mocks/*/': 'KEBAB_CASE',
+      },
+    ],
+    'react/no-unescaped-entities': [
+      'warn',
+      {
+        forbid: [
+          {
+            char: "'",
+            alternatives: [''', '''],
+          },
+          {
+            char: '"',
+            alternatives: ['"', '"'],
+          },
+        ],
+      },
+    ],
+  },
 };

web/.umirc.ts

@@ -3,6 +3,7 @@ import TerserPlugin from 'terser-webpack-plugin';
 import { defineConfig } from 'umi';
 import { appName } from './src/conf.json';
 import routes from './src/routes';
+const ESLintPlugin = require('eslint-webpack-plugin');
 
 export default defineConfig({
   title: appName,
@@ -52,6 +53,15 @@ export default defineConfig({
 
     memo.optimization.minimizer('terser').use(TerserPlugin); // Fixed the issue that the page displayed an error after packaging lexical with terser
 
+    // memo.plugin('eslint').use(ESLintPlugin, [
+    //   {
+    //     extensions: ['js', 'ts', 'tsx'],
+    //     failOnError: true,
+    //     exclude: ['**/node_modules/**', '**/mfsu**', '**/mfsu-virtual-entry**'],
+    //     files: ['src/**/*.{js,ts,tsx}'],
+    //   },
+    // ]);
+
     return memo;
   },
   tailwindcss: {},

web/package-lock.json

@@ -121,6 +121,8 @@
         "@umijs/plugins": "^4.1.0",
         "@welldone-software/why-did-you-render": "^8.0.3",
         "cross-env": "^7.0.3",
+        "eslint-plugin-check-file": "^2.8.0",
+        "eslint-webpack-plugin": "^4.1.0",
         "html-loader": "^5.1.0",
         "husky": "^9.0.11",
         "jest": "^29.7.0",
@@ -9147,9 +9149,10 @@
       }
     },
     "node_modules/@types/eslint": {
-      "version": "8.56.1",
-      "resolved": "https://registry.npmmirror.com/@types/eslint/-/eslint-8.56.1.tgz",
-      "integrity": "sha512-18PLWRzhy9glDQp3+wOgfLYRWlhgX0azxgJ63rdpoUHyrC9z0f5CkFburjQx4uD7ZCruw85ZtMt6K+L+R8fLJQ==",
+      "version": "9.6.1",
+      "resolved": "https://registry.npmmirror.com/@types/eslint/-/eslint-9.6.1.tgz",
+      "integrity": "sha512-FXx2pKgId/WyYo2jXw63kk7/+TY7u7AziEJxJAnSFzHlqTAS3Ync6SvgYAN/k4/PQpnnVuzoMuVnByKK2qp0ag==",
+      "license": "MIT",
       "peer": true,
       "dependencies": {
         "@types/estree": "*",
@@ -16334,6 +16337,27 @@
         "node": "^12.22.0 || ^14.17.0 || >=16.0.0"
       }
     },
+    "node_modules/eslint-plugin-check-file": {
+      "version": "2.8.0",
+      "resolved": "https://registry.npmmirror.com/eslint-plugin-check-file/-/eslint-plugin-check-file-2.8.0.tgz",
+      "integrity": "sha512-FvvafMTam2WJYH9uj+FuMxQ1y+7jY3Z6P9T4j2214cH0FBxNzTcmeCiGTj1Lxp3mI6kbbgsXvmgewvf+llKYyw==",
+      "dev": true,
+      "license": "Apache-2.0",
+      "dependencies": {
+        "is-glob": "^4.0.3",
+        "micromatch": "^4.0.5"
+      },
+      "engines": {
+        "node": ">=18"
+      },
+      "funding": {
+        "type": "ko_fi",
+        "url": "https://ko-fi.com/huanluo"
+      },
+      "peerDependencies": {
+        "eslint": ">=7.28.0"
+      }
+    },
     "node_modules/eslint-plugin-jest": {
       "version": "27.2.3",
       "resolved": "https://registry.npmmirror.com/eslint-plugin-jest/-/eslint-plugin-jest-27.2.3.tgz",
@@ -16437,6 +16461,141 @@
         "node": ">=10"
       }
     },
+    "node_modules/eslint-webpack-plugin": {
+      "version": "4.1.0",
+      "resolved": "https://registry.npmmirror.com/eslint-webpack-plugin/-/eslint-webpack-plugin-4.1.0.tgz",
+      "integrity": "sha512-C3wAG2jyockIhN0YRLuKieKj2nx/gnE/VHmoHemD5ifnAtY6ZU+jNPfzPoX4Zd6RIbUyWTiZUh/ofUlBhoAX7w==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/eslint": "^8.56.5",
+        "jest-worker": "^29.7.0",
+        "micromatch": "^4.0.5",
+        "normalize-path": "^3.0.0",
+        "schema-utils": "^4.2.0"
+      },
+      "engines": {
+        "node": ">= 14.15.0"
+      },
+      "funding": {
+        "type": "opencollective",
+        "url": "https://opencollective.com/webpack"
+      },
+      "peerDependencies": {
+        "eslint": "^8.0.0",
+        "webpack": "^5.0.0"
+      }
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/@types/eslint": {
+      "version": "8.56.12",
+      "resolved": "https://registry.npmmirror.com/@types/eslint/-/eslint-8.56.12.tgz",
+      "integrity": "sha512-03ruubjWyOHlmljCVoxSuNDdmfZDzsrrz0P2LeJsOXr+ZwFQ+0yQIwNCwt/GYhV7Z31fgtXJTAEs+FYlEL851g==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/estree": "*",
+        "@types/json-schema": "*"
+      }
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/ajv": {
+      "version": "8.17.1",
+      "resolved": "https://registry.npmmirror.com/ajv/-/ajv-8.17.1.tgz",
+      "integrity": "sha512-B/gBuNg5SiMTrPkC+A2+cW0RszwxYmn6VYxB/inlBStS5nx6xHIt/ehKRhIMhqusl7a8LjQoZnjCs5vhwxOQ1g==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "fast-deep-equal": "^3.1.3",
+        "fast-uri": "^3.0.1",
+        "json-schema-traverse": "^1.0.0",
+        "require-from-string": "^2.0.2"
+      },
+      "funding": {
+        "type": "github",
+        "url": "https://github.com/sponsors/epoberezkin"
+      }
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/ajv-keywords": {
+      "version": "5.1.0",
+      "resolved": "https://registry.npmmirror.com/ajv-keywords/-/ajv-keywords-5.1.0.tgz",
+      "integrity": "sha512-YCS/JNFAUyr5vAuhk1DWm1CBxRHW9LbJ2ozWeemrIqpbsqKjHVxYPyi5GC0rjZIT5JxJ3virVTS8wk4i/Z+krw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "fast-deep-equal": "^3.1.3"
+      },
+      "peerDependencies": {
+        "ajv": "^8.8.2"
+      }
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/has-flag": {
+      "version": "4.0.0",
+      "resolved": "https://registry.npmmirror.com/has-flag/-/has-flag-4.0.0.tgz",
+      "integrity": "sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=8"
+      }
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/jest-worker": {
+      "version": "29.7.0",
+      "resolved": "https://registry.npmmirror.com/jest-worker/-/jest-worker-29.7.0.tgz",
+      "integrity": "sha512-eIz2msL/EzL9UFTFFx7jBTkeZfku0yUAyZZZmJ93H2TYEiroIx2PQjEXcwYtYl8zXCxb+PAmA2hLIt/6ZEkPHw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/node": "*",
+        "jest-util": "^29.7.0",
+        "merge-stream": "^2.0.0",
+        "supports-color": "^8.0.0"
+      },
+      "engines": {
+        "node": "^14.15.0 || ^16.10.0 || >=18.0.0"
+      }
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/json-schema-traverse": {
+      "version": "1.0.0",
+      "resolved": "https://registry.npmmirror.com/json-schema-traverse/-/json-schema-traverse-1.0.0.tgz",
+      "integrity": "sha512-NM8/P9n3XjXhIZn1lLhkFaACTOURQXjWhV4BA/RnOv8xvgqtqpAX9IO4mRQxSx1Rlo4tqzeqb0sOlruaOy3dug==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/schema-utils": {
+      "version": "4.3.2",
+      "resolved": "https://registry.npmmirror.com/schema-utils/-/schema-utils-4.3.2.tgz",
+      "integrity": "sha512-Gn/JaSk/Mt9gYubxTtSn/QCV4em9mpAPiR1rqy/Ocu19u/G9J5WWdNoUT4SiV6mFC3y6cxyFcFwdzPM3FgxGAQ==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/json-schema": "^7.0.9",
+        "ajv": "^8.9.0",
+        "ajv-formats": "^2.1.1",
+        "ajv-keywords": "^5.1.0"
+      },
+      "engines": {
+        "node": ">= 10.13.0"
+      },
+      "funding": {
+        "type": "opencollective",
+        "url": "https://opencollective.com/webpack"
+      }
+    },
+    "node_modules/eslint-webpack-plugin/node_modules/supports-color": {
+      "version": "8.1.1",
+      "resolved": "https://registry.npmmirror.com/supports-color/-/supports-color-8.1.1.tgz",
+      "integrity": "sha512-MpUEN2OodtUzxvKQl72cUF7RQ5EiHsGvSsVG0ia9c5RbWGL2CI4C7EpPS8UTBIplnlzZiNuV56w+FuNxy3ty2Q==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "has-flag": "^4.0.0"
+      },
+      "engines": {
+        "node": ">=10"
+      },
+      "funding": {
+        "url": "https://github.com/chalk/supports-color?sponsor=1"
+      }
+    },
     "node_modules/eslint/node_modules/ansi-styles": {
       "version": "4.3.0",
       "resolved": "https://registry.npmmirror.com/ansi-styles/-/ansi-styles-4.3.0.tgz",
@@ -17222,9 +17381,10 @@
       }
     },
     "node_modules/flatted": {
-      "version": "3.2.9",
-      "resolved": "https://registry.npmmirror.com/flatted/-/flatted-3.2.9.tgz",
-      "integrity": "sha512-36yxDn5H7OFZQla0/jFJmbIKTdZAQHngCedGxiMmpNfEZM0sdEeT+WczLQrjK6D7o2aiyLYDnkw0R3JK0Qv1RQ==",
+      "version": "3.3.3",
+      "resolved": "https://registry.npmmirror.com/flatted/-/flatted-3.3.3.tgz",
+      "integrity": "sha512-GX+ysw4PBCz0PzosHDepZGANEuFCMLrnRTiEy9McGjmkCQYwRq4A/X786G/fjM/+OjsWSU1ZrY5qyARZmO/uwg==",
+      "license": "ISC",
       "peer": true
     },
     "node_modules/flatten": {
@@ -24121,9 +24281,10 @@
       "integrity": "sha512-oNh6S2WMHWRZrmutsRmDDfkzKtxF+bc2VxLC9dvtrDIRFln627VsFP6fLMgTryGDljgLPjkrzQSDcPrjPyDJ5w=="
     },
     "node_modules/micromatch": {
-      "version": "4.0.7",
-      "resolved": "https://registry.npmmirror.com/micromatch/-/micromatch-4.0.7.tgz",
-      "integrity": "sha512-LPP/3KorzCwBxfeUuZmaR6bG2kdeHSbe0P2tY3FLRU4vYrjYz5hI4QZwV0njUx3jeuKe67YukQ1LSPZBKDqO/Q==",
+      "version": "4.0.8",
+      "resolved": "https://registry.npmmirror.com/micromatch/-/micromatch-4.0.8.tgz",
+      "integrity": "sha512-PXwfBhYu0hBCPw8Dn0E+WDYb7af3dSLVWKi3HGv84IdF4TyFoC0ysxFd0Goxw7nSv4T/PzEJQxsYsEiFCKo2BA==",
+      "license": "MIT",
       "dependencies": {
         "braces": "^3.0.3",
         "picomatch": "^2.3.1"

web/package.json

@@ -132,6 +132,8 @@
     "@umijs/plugins": "^4.1.0",
     "@welldone-software/why-did-you-render": "^8.0.3",
     "cross-env": "^7.0.3",
+    "eslint-plugin-check-file": "^2.8.0",
+    "eslint-webpack-plugin": "^4.1.0",
     "html-loader": "^5.1.0",
     "husky": "^9.0.11",
     "jest": "^29.7.0",

web/src/components/chunk-method-dialog/index.tsx

@@ -80,7 +80,6 @@ export function ChunkMethodDialog({
   hideModal,
   onOk,
   parserId,
-  documentId,
   documentExtension,
   visible,
   parserConfig,

web/src/components/editable-cell.tsx

@@ -1,4 +1,5 @@
 import { Form, FormInstance, Input, InputRef, Typography } from 'antd';
+import { omit } from 'lodash';
 import React, { useContext, useEffect, useRef, useState } from 'react';
 
 const EditableContext = React.createContext<FormInstance<any> | null>(null);
@@ -15,15 +16,12 @@ interface Item {
   address: string;
 }
 
-export const EditableRow: React.FC<EditableRowProps> = ({
-  index,
-  ...props
-}) => {
+export const EditableRow: React.FC<EditableRowProps> = ({ ...props }) => {
   const [form] = Form.useForm();
   return (
     <Form form={form} component={false}>
       <EditableContext.Provider value={form}>
-        <tr {...props} />
+        <tr {...omit(props, 'index')} />
       </EditableContext.Provider>
     </Form>
   );

web/src/components/embed-container.tsx

@@ -0,0 +1,48 @@
+import { useFetchAppConf } from '@/hooks/logic-hooks';
+import { RefreshCcw } from 'lucide-react';
+import { PropsWithChildren } from 'react';
+import { RAGFlowAvatar } from './ragflow-avatar';
+import { Button } from './ui/button';
+
+type EmbedContainerProps = {
+  title: string;
+  avatar?: string;
+  handleReset?(): void;
+} & PropsWithChildren;
+
+export function EmbedContainer({
+  title,
+  avatar,
+  children,
+  handleReset,
+}: EmbedContainerProps) {
+  const appConf = useFetchAppConf();
+
+  return (
+    <section className="h-[100vh] flex justify-center items-center">
+      <div className="w-40 flex gap-2 absolute left-3 top-12 items-center">
+        <img src="/logo.svg" alt="" />
+        <span className="text-2xl font-bold">{appConf.appName}</span>
+      </div>
+      <div className=" w-[80vw] border rounded-lg">
+        <div className="flex justify-between items-center border-b p-3">
+          <div className="flex gap-2 items-center">
+            <RAGFlowAvatar avatar={avatar} name={title} isPerson />
+            <div className="text-xl text-foreground">{title}</div>
+          </div>
+          <Button
+            variant={'secondary'}
+            className="text-sm text-foreground cursor-pointer"
+            onClick={handleReset}
+          >
+            <div className="flex gap-1 items-center">
+              <RefreshCcw size={14} />
+              <span className="text-lg ">Reset</span>
+            </div>
+          </Button>
+        </div>
+        {children}
+      </div>
+    </section>
+  );
+}

web/src/components/embed-dialog/index.tsx

@@ -23,6 +23,7 @@ import {
 } from '@/constants/common';
 import { useTranslate } from '@/hooks/common-hooks';
 import { IModalProps } from '@/interfaces/common';
+import { Routes } from '@/routes';
 import { zodResolver } from '@hookform/resolvers/zod';
 import { memo, useCallback, useMemo } from 'react';
 import { useForm, useWatch } from 'react-hook-form';
@@ -68,7 +69,7 @@ function EmbedDialog({
 
   const generateIframeSrc = useCallback(() => {
     const { visibleAvatar, locale } = values;
-    let src = `${location.origin}/next-chat/share?shared_id=${token}&from=${from}&auth=${beta}`;
+    let src = `${location.origin}${from === SharedFrom.Agent ? Routes.AgentShare : Routes.ChatShare}?shared_id=${token}&from=${from}&auth=${beta}`;
     if (visibleAvatar) {
       src += '&visible_avatar=1';
     }

web/src/components/embed-dialog/use-show-embed-dialog.ts

@@ -0,0 +1,87 @@
+import { useSetModalState, useTranslate } from '@/hooks/common-hooks';
+import { useFetchManualSystemTokenList } from '@/hooks/user-setting-hooks';
+import { useCallback } from 'react';
+import message from '../ui/message';
+
+export const useShowTokenEmptyError = () => {
+  const { t } = useTranslate('chat');
+
+  const showTokenEmptyError = useCallback(() => {
+    message.error(t('tokenError'));
+  }, [t]);
+  return { showTokenEmptyError };
+};
+
+export const useShowBetaEmptyError = () => {
+  const { t } = useTranslate('chat');
+
+  const showBetaEmptyError = useCallback(() => {
+    message.error(t('betaError'));
+  }, [t]);
+  return { showBetaEmptyError };
+};
+
+export const useFetchTokenListBeforeOtherStep = () => {
+  const { showTokenEmptyError } = useShowTokenEmptyError();
+  const { showBetaEmptyError } = useShowBetaEmptyError();
+
+  const { data: tokenList, fetchSystemTokenList } =
+    useFetchManualSystemTokenList();
+
+  let token = '',
+    beta = '';
+
+  if (Array.isArray(tokenList) && tokenList.length > 0) {
+    token = tokenList[0].token;
+    beta = tokenList[0].beta;
+  }
+
+  token =
+    Array.isArray(tokenList) && tokenList.length > 0 ? tokenList[0].token : '';
+
+  const handleOperate = useCallback(async () => {
+    const ret = await fetchSystemTokenList();
+    const list = ret;
+    if (Array.isArray(list) && list.length > 0) {
+      if (!list[0].beta) {
+        showBetaEmptyError();
+        return false;
+      }
+      return list[0]?.token;
+    } else {
+      showTokenEmptyError();
+      return false;
+    }
+  }, [fetchSystemTokenList, showBetaEmptyError, showTokenEmptyError]);
+
+  return {
+    token,
+    beta,
+    handleOperate,
+  };
+};
+
+export const useShowEmbedModal = () => {
+  const {
+    visible: embedVisible,
+    hideModal: hideEmbedModal,
+    showModal: showEmbedModal,
+  } = useSetModalState();
+
+  const { handleOperate, token, beta } = useFetchTokenListBeforeOtherStep();
+
+  const handleShowEmbedModal = useCallback(async () => {
+    const succeed = await handleOperate();
+    if (succeed) {
+      showEmbedModal();
+    }
+  }, [handleOperate, showEmbedModal]);
+
+  return {
+    showEmbedModal: handleShowEmbedModal,
+    hideEmbedModal,
+    embedVisible,
+    embedToken: token,
+    beta,
+  };
+};

web/src/components/highlight-markdown/index.tsx

@@ -31,7 +31,7 @@ const HightLightMarkdown = ({
       components={
         {
           code(props: any) {
-            const { children, className, node, ...rest } = props;
+            const { children, className, ...rest } = props;
             const match = /language-(\w+)/.exec(className || '');
             return match ? (
               <SyntaxHighlighter

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

@@ -70,6 +70,10 @@ const KnowledgeBaseItem = ({
 
 export default KnowledgeBaseItem;
 
+function buildQueryVariableOptionsByShowVariable(showVariable?: boolean) {
+  return showVariable ? useBuildQueryVariableOptions : () => [];
+}
+
 export function KnowledgeBaseFormField({
   showVariable = false,
 }: {
@@ -84,7 +88,7 @@ export function KnowledgeBaseFormField({
     (x) => x.parser_id !== DocumentParserType.Tag,
   );
 
-  const nextOptions = useBuildQueryVariableOptions();
+  const nextOptions = buildQueryVariableOptionsByShowVariable(showVariable)();
 
   const knowledgeOptions = filteredKnowledgeList.map((x) => ({
     label: x.name,

web/src/components/large-model-form-field.tsx

@@ -16,7 +16,7 @@ import { Funnel } from 'lucide-react';
 import { useFormContext, useWatch } from 'react-hook-form';
 import { useTranslation } from 'react-i18next';
 import { z } from 'zod';
-import { NextLLMSelect } from './llm-select/next';
+import { NextInnerLLMSelectProps, NextLLMSelect } from './llm-select/next';
 import { Button } from './ui/button';
 
 const ModelTypes = [
@@ -38,7 +38,13 @@ export const LargeModelFilterFormSchema = {
   llm_filter: z.string().optional(),
 };
 
-export function LargeModelFormField() {
+type LargeModelFormFieldProps = Pick<
+  NextInnerLLMSelectProps,
+  'showSpeech2TextModel'
+>;
+export function LargeModelFormField({
+  showSpeech2TextModel: showTTSModel,
+}: LargeModelFormFieldProps) {
   const form = useFormContext();
   const { t } = useTranslation();
   const filter = useWatch({ control: form.control, name: 'llm_filter' });
@@ -85,7 +91,11 @@ export function LargeModelFormField() {
               />
 
               <FormControl>
-                <NextLLMSelect {...field} filter={filter} />
+                <NextLLMSelect
+                  {...field}
+                  filter={filter}
+                  showSpeech2TextModel={showTTSModel}
+                />
               </FormControl>
             </section>
 
@@ -96,3 +106,22 @@ export function LargeModelFormField() {
     </>
   );
 }
+
+export function LargeModelFormFieldWithoutFilter() {
+  const form = useFormContext();
+
+  return (
+    <FormField
+      control={form.control}
+      name="llm_id"
+      render={({ field }) => (
+        <FormItem>
+          <FormControl>
+            <NextLLMSelect {...field} />
+          </FormControl>
+          <FormMessage />
+        </FormItem>
+      )}
+    />
+  );
+}