Merge branch 'develop' into patch-2

Author: WMC001

backend/apps/prompt_app.py

@@ -27,7 +27,9 @@ async def generate_and_save_system_prompt_api(
             task_description=prompt_request.task_description,
             user_id=user_id,
             tenant_id=tenant_id,
-            language=language
+            language=language,
+            tool_ids=prompt_request.tool_ids,
+            sub_agent_ids=prompt_request.sub_agent_ids
         ), media_type="text/event-stream")
     except Exception as e:
         logger.exception(f"Error occurred while generating system prompt: {e}")

backend/apps/vectordatabase_app.py

@@ -3,6 +3,7 @@
 from typing import Any, Dict, List, Optional
 
 from fastapi import APIRouter, Body, Depends, Header, HTTPException, Path, Query
+from fastapi.responses import JSONResponse
 
 from consts.model import IndexingResponse
 from nexent.vector_database.base import VectorDatabaseCore
@@ -195,3 +196,33 @@ def health_check(vdb_core: VectorDatabaseCore = Depends(get_vector_db_core)):
         return ElasticSearchService.health_check(vdb_core)
     except Exception as e:
         raise HTTPException(status_code=HTTPStatus.INTERNAL_SERVER_ERROR, detail=f"{str(e)}")
+
+
+@router.post("/{index_name}/chunks")
+def get_index_chunks(
+        index_name: str = Path(...,
+                               description="Name of the index to get chunks from"),
+        page: int = Query(
+            None, description="Page number (1-based) for pagination"),
+        page_size: int = Query(
+            None, description="Number of records per page for pagination"),
+        path_or_url: Optional[str] = Query(
+            None, description="Filter chunks by document path_or_url"),
+        vdb_core: VectorDatabaseCore = Depends(get_vector_db_core)
+):
+    """Get chunks from the specified index, with optional pagination support"""
+    try:
+        result = ElasticSearchService.get_index_chunks(
+            index_name=index_name,
+            page=page,
+            page_size=page_size,
+            path_or_url=path_or_url,
+            vdb_core=vdb_core,
+        )
+        return JSONResponse(status_code=HTTPStatus.OK, content=result)
+    except Exception as e:
+        error_msg = str(e)
+        logger.error(
+            f"Error getting chunks for index '{index_name}': {error_msg}")
+        raise HTTPException(
+            status_code=HTTPStatus.INTERNAL_SERVER_ERROR, detail=f"Error getting chunks: {error_msg}")

backend/consts/const.py

@@ -279,7 +279,7 @@ class VectorDatabaseType(str, Enum):
     os.getenv("LLM_SLOW_TOKEN_RATE_THRESHOLD", "10.0"))  # tokens per second
 
 # APP Version
-APP_VERSION = "v1.7.5.3"
+APP_VERSION = "v1.7.6"
 
 DEFAULT_ZH_TITLE = "新对话"
 DEFAULT_EN_TITLE = "New Conversation"

backend/consts/model.py

@@ -193,6 +193,8 @@ class GeneratePromptRequest(BaseModel):
     task_description: str
     agent_id: int
     model_id: int
+    tool_ids: Optional[List[int]] = None  # Optional: tool IDs from frontend (takes precedence over database query)
+    sub_agent_ids: Optional[List[int]] = None  # Optional: sub-agent IDs from frontend (takes precedence over database query)
 
 
 class GenerateTitleRequest(BaseModel):

backend/services/prompt_service.py

@@ -2,17 +2,17 @@
 import logging
 import queue
 import threading
+from typing import Optional, List
 
 from jinja2 import StrictUndefined, Template
 from smolagents import OpenAIServerModel
 
-from consts.const import LANGUAGE, MODEL_CONFIG_MAPPING, MESSAGE_ROLE, THINK_END_PATTERN, THINK_START_PATTERN
+from consts.const import LANGUAGE, MESSAGE_ROLE, THINK_END_PATTERN, THINK_START_PATTERN
 from consts.model import AgentInfoRequest
-from database.agent_db import update_agent, query_sub_agents_id_list, search_agent_info_by_agent_id
+from database.agent_db import update_agent, search_agent_info_by_agent_id
 from database.model_management_db import get_model_by_model_id
 from database.tool_db import query_tools_by_ids
-from services.agent_service import get_enable_tool_id_by_agent_id
-from utils.config_utils import tenant_config_manager, get_model_name_from_config
+from utils.config_utils import get_model_name_from_config
 from utils.prompt_template_utils import get_prompt_generate_prompt_template
 
 # Configure logging
@@ -34,7 +34,7 @@ def _process_thinking_tokens(new_token: str, is_thinking: bool, token_join: list
     """
     # Handle thinking mode
     if is_thinking:
-        return not (THINK_END_PATTERN in new_token)
+        return THINK_END_PATTERN not in new_token
 
     # Handle start of thinking
     if THINK_START_PATTERN in new_token:
@@ -98,14 +98,16 @@ def call_llm_for_system_prompt(model_id: int, user_prompt: str, system_prompt: s
         raise e
 
 
-def gen_system_prompt_streamable(agent_id: int, model_id: int, task_description: str, user_id: str, tenant_id: str, language: str):
+def gen_system_prompt_streamable(agent_id: int, model_id: int, task_description: str, user_id: str, tenant_id: str, language: str, tool_ids: Optional[List[int]] = None, sub_agent_ids: Optional[List[int]] = None):
     for system_prompt in generate_and_save_system_prompt_impl(
         agent_id=agent_id,
         model_id=model_id,
         task_description=task_description,
         user_id=user_id,
         tenant_id=tenant_id,
-        language=language
+        language=language,
+        tool_ids=tool_ids,
+        sub_agent_ids=sub_agent_ids
     ):
         # SSE format, each message ends with \n\n
         yield f"data: {json.dumps({'success': True, 'data': system_prompt}, ensure_ascii=False)}\n\n"
@@ -116,17 +118,35 @@ def generate_and_save_system_prompt_impl(agent_id: int,
                                          task_description: str,
                                          user_id: str,
                                          tenant_id: str,
-                                         language: str):
-    # Get description of tool and agent
-    # In create mode (agent_id=0), return empty lists
-    if agent_id == 0:
+                                         language: str,
+                                         tool_ids: Optional[List[int]] = None,
+                                         sub_agent_ids: Optional[List[int]] = None):
+    # Get description of tool and agent from frontend-provided IDs
+    # Frontend always provides tool_ids and sub_agent_ids (could be empty arrays)
+
+    # Handle tool IDs
+    if tool_ids and len(tool_ids) > 0:
+        tool_info_list = query_tools_by_ids(tool_ids)
+        logger.debug(f"Using frontend-provided tool IDs: {tool_ids}")
+    else:
         tool_info_list = []
+        logger.debug("No tools selected (empty tool_ids list)")
+
+    # Handle sub-agent IDs
+    if sub_agent_ids and len(sub_agent_ids) > 0:
         sub_agent_info_list = []
+        for sub_agent_id in sub_agent_ids:
+            try:
+                sub_agent_info = search_agent_info_by_agent_id(
+                    agent_id=sub_agent_id, tenant_id=tenant_id)
+                sub_agent_info_list.append(sub_agent_info)
+            except Exception as e:
+                logger.warning(
+                    f"Failed to get sub-agent info for agent_id {sub_agent_id}: {str(e)}")
+        logger.debug(f"Using frontend-provided sub-agent IDs: {sub_agent_ids}")
     else:
-        tool_info_list = get_enabled_tool_description_for_generate_prompt(
-            tenant_id=tenant_id, agent_id=agent_id)
-        sub_agent_info_list = get_enabled_sub_agent_description_for_generate_prompt(
-            tenant_id=tenant_id, agent_id=agent_id)
+        sub_agent_info_list = []
+        logger.debug("No sub-agents selected (empty sub_agent_ids list)")
 
     # 1. Real-time streaming push
     final_results = {"duty": "", "constraint": "", "few_shots": "", "agent_var_name": "", "agent_display_name": "",
@@ -292,27 +312,3 @@ def join_info_for_generate_system_prompt(prompt_for_generate, sub_agent_info_lis
         "assistant_description": assistant_description
     })
     return content
-
-
-def get_enabled_tool_description_for_generate_prompt(agent_id: int, tenant_id: str):
-    # Get tool information
-    logger.info("Fetching tool instances")
-    tool_id_list = get_enable_tool_id_by_agent_id(
-        agent_id=agent_id, tenant_id=tenant_id)
-    tool_info_list = query_tools_by_ids(tool_id_list)
-    return tool_info_list
-
-
-def get_enabled_sub_agent_description_for_generate_prompt(agent_id: int, tenant_id: str):
-    logger.info("Fetching sub-agents information")
-
-    sub_agent_id_list = query_sub_agents_id_list(
-        main_agent_id=agent_id, tenant_id=tenant_id)
-
-    sub_agent_info_list = []
-    for sub_agent_id in sub_agent_id_list:
-        sub_agent_info = search_agent_info_by_agent_id(
-            agent_id=sub_agent_id, tenant_id=tenant_id)
-
-        sub_agent_info_list.append(sub_agent_info)
-    return sub_agent_info_list

backend/services/vectordatabase_service.py

@@ -35,6 +35,9 @@
 from utils.config_utils import tenant_config_manager, get_model_name_from_config
 from utils.file_management_utils import get_all_files_status, get_file_size
 
+ALLOWED_CHUNK_FIELDS = {"filename",
+                        "path_or_url", "content", "create_time", "id"}
+
 # Configure logging
 logger = logging.getLogger("vectordatabase_service")
 
@@ -572,7 +575,8 @@ async def list_files(
                     'file_size': file_info.get('file_size', 0),
                     'create_time': int(utc_create_timestamp * 1000),
                     'status': "COMPLETED",
-                    'latest_task_id': ''
+                    'latest_task_id': '',
+                    'chunk_count': file_info.get('chunk_count', 0)
                 }
                 files.append(file_data)
 
@@ -630,7 +634,7 @@ async def list_files(
                 # Initialize chunks for all files
                 for file_data in files:
                     file_data['chunks'] = []
-                    file_data['chunk_count'] = 0
+                    file_data['chunk_count'] = file_data.get('chunk_count', 0)
 
                 if msearch_body:
                     try:
@@ -667,7 +671,7 @@ async def list_files(
             else:
                 for file_data in files:
                     file_data['chunks'] = []
-                    file_data['chunk_count'] = 0
+                    file_data['chunk_count'] = file_data.get('chunk_count', 0)
 
             return {"files": files}
 
@@ -919,3 +923,62 @@ def get_summary(index_name: str = Path(..., description="Name of the index to ge
         except Exception as e:
             error_msg = f"Failed to get summary: {str(e)}"
             raise Exception(error_msg)
+
+    @staticmethod
+    def get_index_chunks(
+        index_name: str,
+        page: Optional[int] = None,
+        page_size: Optional[int] = None,
+        path_or_url: Optional[str] = None,
+        vdb_core: VectorDatabaseCore = Depends(get_vector_db_core),
+    ):
+        """
+        Retrieve chunk records for the specified index with optional pagination.
+
+        Args:
+            index_name: Name of the index to query
+            page: Page number (1-based) when paginating
+            page_size: Page size when paginating
+            path_or_url: Optional document filter
+            vdb_core: VectorDatabaseCore instance
+
+        Returns:
+            Dictionary containing status, chunk list, total, and pagination metadata
+        """
+        try:
+            result = vdb_core.get_index_chunks(
+                index_name,
+                page=page,
+                page_size=page_size,
+                path_or_url=path_or_url,
+            )
+            raw_chunks = result.get("chunks", [])
+            total = result.get("total", len(raw_chunks))
+            result_page = result.get("page", page)
+            result_page_size = result.get("page_size", page_size)
+
+            filtered_chunks: List[Any] = []
+            for chunk in raw_chunks:
+                if isinstance(chunk, dict):
+                    filtered_chunks.append(
+                        {
+                            field: chunk.get(field)
+                            for field in ALLOWED_CHUNK_FIELDS
+                            if field in chunk
+                        }
+                    )
+                else:
+                    filtered_chunks.append(chunk)
+
+            return {
+                "status": "success",
+                "message": f"Successfully retrieved {len(filtered_chunks)} chunks from index {index_name}",
+                "chunks": filtered_chunks,
+                "total": total,
+                "page": result_page,
+                "page_size": result_page_size
+            }
+        except Exception as e:
+            error_msg = f"Error retrieving chunks from index {index_name}: {str(e)}"
+            logger.error(error_msg)
+            raise Exception(error_msg)

doc/docs/zh/opensource-memorial-wall.md

@@ -15,6 +15,9 @@
 每条消息应包含您的姓名/昵称和日期。
 请保持消息的礼貌和尊重，符合我们的行为准则。
 -->
+::: china-king-hs - 2025-11-20
+希望能正常使用nexent
+:::
 
 ::: info happyzhang - 2025-11-13
 也许我们正见证着未来的“后起之秀”😀
@@ -434,6 +437,10 @@ Nexent功能如此之强大，给我很多帮助，感谢开发者！厉害
 感谢 Nexent 让我踏上了开源之旅！给我一个机会制作智能体
 :::
 
-::: info chengyudan - 2025-10-14 
+::: info chengyudan - 2025-10-20 
 感谢 Nexent 让我踏上了开源之旅！ 
 :::
+
+::: info user - 2025-11-20
+学习ai - agent非常好的项目，后面会持续输出贡献！
+:::

docker/.env.example

@@ -1,4 +1,4 @@
-# ===== Necessary Configs (Neccessary till now, will be migrated to frontend page) =====
+# ===== Necessary Configs (Neccessary till now, will be migrated to frontend page) =====
 
 # Voice Service Config
 APPID=app_id

frontend/app/[locale]/agents/AgentConfiguration.tsx

@@ -161,12 +161,25 @@ export default forwardRef<AgentConfigHandle, AgentConfigProps>(function AgentCon
       const currentAgentName = agentName;
       const currentAgentDisplayName = agentDisplayName;
 
+      // Extract tool IDs from selected tools (convert string IDs to numbers)
+      // Always pass tool_ids array (empty array means no tools selected, undefined means use database)
+      // In edit mode, we want to use current selection, so pass the array even if empty
+      const toolIds = selectedTools.map((tool) => Number(tool.id));
+      
+      // Get sub-agent IDs from enabledAgentIds
+      // Always pass sub_agent_ids array (empty array means no sub-agents selected, undefined means use database)
+      // In edit mode, we want to use current selection, so pass the array even if empty
+      const subAgentIds = [...enabledAgentIds];
+
       // Call backend API to generate agent prompt
+      // Pass tool_ids and sub_agent_ids to use frontend selection instead of database query
       await generatePromptStream(
         {
           agent_id: agentIdToUse,
           task_description: businessLogic,
           model_id: selectedModel?.id?.toString() || "",
+          tool_ids: toolIds,
+          sub_agent_ids: subAgentIds,
         },
         (data) => {
           // Process streaming response data