auto-coder 0.1.292__tar.gz → 0.1.294__tar.gz

{auto_coder-0.1.292 → auto_coder-0.1.294}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.292
+Version: 0.1.294
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/auto_coder.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.292
+Version: 0.1.294
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/auto_coder.egg-info/SOURCES.txt

@@ -92,6 +92,7 @@ src/autocoder/common/text.py
 src/autocoder/common/types.py
 src/autocoder/common/utils_code_auto_generate.py
 src/autocoder/common/mcp_servers/__init__.py
+src/autocoder/common/mcp_servers/mcp_server_gpt4o_mini_search.py
 src/autocoder/common/mcp_servers/mcp_server_perplexity.py
 src/autocoder/data/byzerllm.md
 src/autocoder/data/tokenizer.json

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/auto_coder_runner.py

@@ -39,7 +39,7 @@ import git
 from autocoder.common import git_utils
 from autocoder.chat_auto_coder_lang import get_message
 from autocoder.agent.auto_guess_query import AutoGuessQuery
-from autocoder.common.mcp_server import get_mcp_server, McpRequest, McpInstallRequest, McpRemoveRequest, McpListRequest, McpListRunningRequest, McpRefreshRequest
+from autocoder.common.mcp_server import get_mcp_server, McpRequest, McpInstallRequest, McpRemoveRequest, McpListRequest, McpListRunningRequest, McpRefreshRequest,McpServerInfoRequest
 import byzerllm
 from byzerllm.utils import format_str_jinja2
 from autocoder.common.memory_manager import get_global_memory_file_paths 
@@ -1133,7 +1133,7 @@ def mcp(query: str):
             printer.print_in_terminal("mcp_list_builtin_title")
             printer.print_str_in_terminal(response.result)
         return
-
+        
     # Handle refresh command
     if query.startswith("/refresh"):
         server_name = query.replace("/refresh", "", 1).strip()    
@@ -1183,7 +1183,21 @@ def mcp(query: str):
         if os.path.exists(temp_yaml):
             os.remove(temp_yaml)
 
-    mcp_server = get_mcp_server()    
+    mcp_server = get_mcp_server()   
+
+
+    if query.startswith("/info"):
+        response = mcp_server.send_request(McpServerInfoRequest(
+            model=args.inference_model or args.model,
+            product_mode=args.product_mode
+        ))
+        if response.error:
+            printer.print_in_terminal("mcp_server_info_error", style="red", error=response.error)
+        else:
+            printer.print_in_terminal("mcp_server_info_title")
+            printer.print_str_in_terminal(response.result)
+        return
+
     response = mcp_server.send_request(
         McpRequest(
             query=query,

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/auto_coder_server.py

@@ -34,6 +34,14 @@ import sys
 import io
 from autocoder.utils.log_capture import LogCapture
 
+# If support dotenv, use it
+if os.path.exists(".env"):
+    try:
+        from dotenv import load_dotenv
+        load_dotenv()
+    except ImportError:
+        pass
+
 def convert_yaml_config_to_str(yaml_config):
     yaml_content = yaml.safe_dump(
         yaml_config,

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/chat_auto_coder_lang.py

@@ -138,6 +138,8 @@ MESSAGES = {
         "official_doc": "Official Documentation: https://uelng8wukz.feishu.cn/wiki/NhPNwSRcWimKFIkQINIckloBncI",
         "plugins_desc": "Manage plugins",
         "plugins_usage": "Usage: /plugins <command>\nAvailable subcommands:\n  /plugins /list - List all available plugins\n  /plugins /load <name> - Load a plugin\n  /plugins /unload <name> - Unload a plugin\n  /plugins/dirs - List plugin directories\n  /plugins/dirs /add <path> - Add a plugin directory\n  /plugins/dirs /remove <path> - Remove a plugin directory\n  /plugins/dirs /clear - Clear all plugin directories",
+        "mcp_server_info_error": "Error getting MCP server info: {{ error }}",
+        "mcp_server_info_title": "Connected MCP Server Info",
     },
     "zh": {
         "auto_command_analyzing": "正在分析命令请求",
@@ -275,6 +277,8 @@ MESSAGES = {
         "official_doc": "官方文档: https://uelng8wukz.feishu.cn/wiki/NhPNwSRcWimKFIkQINIckloBncI",
         "plugins_desc": "管理插件",
         "plugins_usage": "用法: /plugins <命令>\n可用的子命令:\n  /plugins /list - 列出所有可用插件\n  /plugins /load <名称> - 加载一个插件\n  /plugins /unload <名称> - 卸载一个插件\n  /plugins/dirs - 列出插件目录\n  /plugins/dirs /add <路径> - 添加一个插件目录\n  /plugins/dirs /remove <路径> - 移除一个插件目录\n  /plugins/dirs /clear - 清除所有插件目录",
+        "mcp_server_info_error": "获取MCP服务器信息时出错: {{ error }}",
+        "mcp_server_info_title": "已连接的MCP服务器信息",
     },
 }
 

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/commands/auto_command.py

@@ -24,6 +24,7 @@ from autocoder.common.global_cancel import global_cancel
 from autocoder.common.auto_configure import config_readme
 from autocoder.utils.auto_project_type import ProjectTypeAnalyzer
 from rich.text import Text
+from autocoder.common.mcp_server import get_mcp_server,McpServerInfoRequest
 
 class CommandMessage(BaseModel):
     role: str
@@ -159,7 +160,17 @@ class CommandAutoTuner:
         self.memory_config = memory_config
         self.command_config = command_config
         self.tools = AutoCommandTools(args=args, llm=self.llm)
-        self.project_type_analyzer = ProjectTypeAnalyzer(args=args, llm=self.llm)        
+        self.project_type_analyzer = ProjectTypeAnalyzer(args=args, llm=self.llm) 
+        try:
+            self.mcp_server = get_mcp_server()
+            mcp_server_info_response = self.mcp_server.send_request(McpServerInfoRequest(
+                model=args.inference_model or args.model,
+                product_mode=args.product_mode
+            ))
+            self.mcp_server_info = mcp_server_info_response.result
+        except Exception as e:
+            logger.error(f"Error getting MCP server info: {str(e)}")
+            self.mcp_server_info = ""
     
     def get_conversations(self) -> List[CommandMessage]:
         """Get conversation history from memory file"""
@@ -1210,12 +1221,26 @@ class CommandAutoTuner:
          response_user(response="你好，我是 auto-coder")
         </usage>
         </command>
-        </commands>
-        
-        
+                
+        <command>
+        <name>execute_mcp_server</name>
+        <description>执行MCP服务器</description>
+        <usage>
+         该函数接受一个参数 query, 为要执行的MCP服务器查询字符串。
+
+         你可以根据下面已经连接的 mcp server 信息，来决定个是否调用该函数，注意该函数会更具你的 query 
+         自动选择合适的 mcp server 来执行。如果你想某个特定的 server 来执行，你可以在 query 中说明你想哪个 server 执行。
+
+         <mcp_server_info>
+         {{ mcp_server_info }}
+         </mcp_server_info>
+         
+        </usage>
+        </command>                        
         '''
         return {
-            "config_readme": config_readme.prompt()
+            "config_readme": config_readme.prompt(),
+            "mcp_server_info": self.mcp_server_info            
         }
 
     def execute_auto_command(self, command: str, parameters: Dict[str, Any]) -> None:
@@ -1257,6 +1282,7 @@ class CommandAutoTuner:
             "read_file_with_keyword_ranges": self.tools.read_file_with_keyword_ranges,
             "get_project_type": self.project_type_analyzer.analyze,
             "response_user": self.tools.response_user,
+            "execute_mcp_server": self.tools.execute_mcp_server,
                                     
         }
 

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/commands/tools.py

@@ -62,6 +62,28 @@ class AutoCommandTools:
         self.result_manager = ResultManager()
         self.printer = Printer()
 
+    def execute_mcp_server(self, query: str) -> str:
+        from autocoder.common.mcp_server import get_mcp_server, McpRequest, McpInstallRequest, McpRemoveRequest, McpListRequest, McpListRunningRequest, McpRefreshRequest                               
+        mcp_server = get_mcp_server()
+        response = mcp_server.send_request(
+            McpRequest(
+                query=query,
+                model=self.args.inference_model or self.args.model,
+                product_mode=self.args.product_mode
+            )
+        )
+        
+        result = response.result
+        
+        self.result_manager.append(content=result, meta = {
+            "action": "execute_mcp_server",
+            "input": {
+                "query": query
+            }
+        })
+        return result
+
+
     def ask_user(self,question:str) -> str:
         '''
         如果你对用户的问题有什么疑问，或者你想从用户收集一些额外信息，可以调用此方法。

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/common/auto_coder_lang.py

@@ -191,7 +191,9 @@ MESSAGES = {
         "super_big_filter_failed": "❌ Super big filter failed: {{ error }}.",
         "super_big_filter_stats": "{{ model_names }} Super big filter completed in {{ elapsed_time }} seconds, input tokens: {{ input_tokens }}, output tokens: {{ output_tokens }}, input cost: {{ input_cost }}, output cost: {{ output_cost }}, speed: {{ speed }} tokens/s, chunk_index: {{ chunk_index }}",
         "super_big_filter_splitting": "⚠️ Index file is extremely large ({{ tokens_len }}/{{ max_tokens }}). The query will be split into {{ split_size }} chunks for processing.",
-        "super_big_filter_title": "{{ model_name }} is analyzing how to filter extremely large context..."
+        "super_big_filter_title": "{{ model_name }} is analyzing how to filter extremely large context...",
+        "mcp_server_info_error": "Error getting MCP server info: {{ error }}",
+        "mcp_server_info_title": "Connected MCP Server Info",
     },
     "zh": {
         "file_sliding_window_processing": "文件 {{ file_path }} 过大 ({{ tokens }} tokens)，正在使用滑动窗口处理...",
@@ -380,7 +382,9 @@ MESSAGES = {
         "super_big_filter_failed": "❌ 超大过滤器失败: {{ error }}.",
         "super_big_filter_stats": "{{ model_names }} 超大过滤器完成耗时 {{ elapsed_time }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 输入成本: {{ input_cost }}, 输出成本: {{ output_cost }}, 速度: {{ speed }} tokens/秒, 块索引: {{ chunk_index }}",
         "super_big_filter_splitting": "⚠️ 索引文件极其庞大 ({{ tokens_len }}/{{ max_tokens }})。查询将被分成 {{ split_size }} 个部分进行处理。",
-        "super_big_filter_title": "{{ model_name }} 正在分析如何过滤极大规模上下文..."
+        "super_big_filter_title": "{{ model_name }} 正在分析如何过滤极大规模上下文...",
+        "mcp_server_info_error": "获取MCP服务器信息时出错: {{ error }}",
+        "mcp_server_info_title": "已连接的MCP服务器信息",
     }}
 
 

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/common/mcp_hub.py

@@ -2,8 +2,12 @@ import os
 import json
 import asyncio
 import aiohttp
+import importlib
+import pkgutil
+import re
+import inspect
 from datetime import datetime, timedelta
-from typing import Dict, List, Optional, Any, Set, Optional
+from typing import Dict, List, Optional, Any, Set, Optional, Tuple
 from pathlib import Path
 from pydantic import BaseModel, Field
 
@@ -62,23 +66,53 @@ class McpConnection:
         self.session = session        
 
 
-MCP_PERPLEXITY_SERVER = '''
-{
-    "perplexity": {
-        "command": "python",
-        "args": [
-            "-m", "autocoder.common.mcp_servers.mcp_server_perplexity"
-        ],
-        "env": {
-            "PERPLEXITY_API_KEY": "{{PERPLEXITY_API_KEY}}"
-        }
-    }
-}
-'''
-
-MCP_BUILD_IN_SERVERS = {
-    "perplexity": json.loads(MCP_PERPLEXITY_SERVER)["perplexity"]
-}
+def _generate_server_configs() -> Tuple[Dict[str, Any], Dict[str, str]]:
+    """
+    Scan the autocoder.common.mcp_servers directory for mcp_server_*.py files
+    and generate server configurations.
+    
+    Returns:
+        Tuple of (built-in servers dict, JSON templates dict)
+    """
+    servers = {}
+    templates = {}
+    
+    try:
+        package_name = "autocoder.common.mcp_servers"
+        package = importlib.import_module(package_name)
+        
+        # Find all modules in the package
+        for _, name, _ in pkgutil.iter_modules(package.__path__, package.__name__ + "."):
+            # Only process modules that start with "mcp_server_"
+            base_name = name.split(".")[-1]
+            if base_name.startswith("mcp_server_"):
+                # Generate a friendly server name
+                friendly_name = base_name[11:]
+                                
+                # Create env dictionary with placeholders
+                env_dict = {}
+                
+                # Create server configuration
+                config = {
+                    "command": "python",
+                    "args": ["-m", name],
+                    "env": env_dict
+                }                                
+                
+                # Store in dictionaries
+                servers[friendly_name] = config
+                templates[friendly_name] = json.dumps({friendly_name: config}, indent=4)
+                
+                logger.info(f"Detected MCP server: {friendly_name}")
+    
+    except Exception as e:
+        logger.error(f"Error generating server configs: {e}")
+    
+    return servers, templates
+
+
+# Automatically generate server configurations
+MCP_BUILD_IN_SERVERS, MCP_SERVER_TEMPLATES = _generate_server_configs()
 
 
 class McpHub:
@@ -422,3 +456,10 @@ class McpHub:
         """
         for name in list(self.connections.keys()):
             await self.delete_connection(name)
+
+    @classmethod
+    def get_server_templates(cls) -> Dict[str, str]:
+        """
+        Get all available server templates as JSON strings
+        """
+        return MCP_SERVER_TEMPLATES

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/common/mcp_server.py

@@ -53,6 +53,12 @@ class McpRefreshRequest:
     """Request to refresh MCP server connections"""
     name: Optional[str] = None
 
+@dataclass
+class McpServerInfoRequest:
+    """Request to get MCP server info"""
+    model: Optional[str] = None
+    product_mode: Optional[str] = None
+
 
 @dataclass
 class McpExternalServer(BaseModel):
@@ -346,6 +352,18 @@ class McpServer:
                         await self._response_queue.put(McpResponse(
                             result="", error=get_message_with_format("mcp_list_builtin_error", error=str(e))))
 
+                elif isinstance(request, McpServerInfoRequest):
+                    try:
+                        llm = get_single_llm(request.model, product_mode=request.product_mode)
+                        mcp_executor = McpExecutor(hub, llm)
+                        result = mcp_executor.get_connected_servers_info()
+                        await self._response_queue.put(McpResponse(result=result))
+                    except Exception as e:
+                        import traceback
+                        traceback.print_exc()
+                        await self._response_queue.put(McpResponse(
+                            result="", error=get_message_with_format("mcp_server_info_error", error=str(e))))
+
                 elif isinstance(request, McpListRunningRequest):
                     try:
                         running_servers = "\n".join(

auto_coder-0.1.294/src/autocoder/common/mcp_servers/mcp_server_gpt4o_mini_search.py

@@ -0,0 +1,153 @@
+from os import getenv
+from textwrap import dedent
+import sys
+
+import mcp.server.stdio
+import mcp.types as types
+from mcp.server import NotificationOptions, Server
+from mcp.server.models import InitializationOptions
+import json
+from openai import OpenAI
+
+OPENAI_API_KEY = getenv("OPENAI_API_KEY")
+# Check if API key is empty or None
+if not OPENAI_API_KEY:
+    print("Error: OPENAI_API_KEY environment variable is not set. Please set it before running this server.", file=sys.stderr)
+    sys.exit(1)
+
+OPENAI_API_BASE_URL = getenv(
+    "OPENAI_API_BASE_URL", "https://api.openai.com/v1")
+
+server = Server("mcp-server-gpt4o-mini-search")
+
+client = OpenAI(
+    api_key=OPENAI_API_KEY,
+    base_url=OPENAI_API_BASE_URL
+)
+
+
+@server.list_tools()
+async def handle_list_tools() -> list[types.Tool]:
+    return [
+        types.Tool(
+            name="gpt4o_mini_search",
+            description=dedent(
+                """
+                GPT-4o mini with search enables agents to gather information from the internet 
+                in real-time, providing up-to-date answers with source citations.
+                This tool is ideal for fact-checking, research, and accessing current information
+                that might not be in the model's training data.
+                
+                The search-enhanced responses include relevant web sources to support the information
+                provided, making it useful for obtaining verified and recent information.
+                
+                [Response structure]
+                - id: A unique identifier for the response
+                - model: The model used (gpt-4o-mini-search-preview)
+                - object: The object type ("chat.completion")
+                - created: The Unix timestamp when the completion was created
+                - choices[]: The list of completion choices generated
+                - usage: Usage statistics for the completion request
+                """
+            ),
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "system_message": {
+                        "type": "string",
+                        "description": "Optional custom system message. If not provided, a default search-optimized system message will be used.",
+                    },
+                    "messages": {
+                        "type": "array",
+                        "description": "A list of messages comprising the conversation so far (excluding system message which is handled separately).",
+                        "items": {
+                            "type": "object",
+                            "properties": {
+                                "content": {
+                                    "type": "string",
+                                    "description": "The contents of the message in this turn of conversation.",
+                                },
+                                "role": {
+                                    "type": "string",
+                                    "description": "The role of the speaker in this turn of conversation.",
+                                    "enum": ["user", "assistant"],
+                                },
+                            },
+                            "required": ["content", "role"],
+                        },
+                    },
+                },
+                "required": ["messages"],
+            },
+        )
+    ]
+
+
+@server.call_tool()
+async def handle_call_tool(
+    name: str, arguments: dict
+) -> list[types.TextContent | types.ImageContent | types.EmbeddedResource]:
+    if name != "gpt4o_mini_search":
+        raise ValueError(f"Unknown tool: {name}")
+
+    # Extract user messages
+    user_messages = arguments.get("messages", [])
+
+    # Define default system message if not provided
+    default_system_message = (
+        "你是专业搜索助手，需要：\n"
+        "1. 提供基于用户查询的清晰格式化信息\n"
+        "2. 使用[标题](URL)格式嵌入链接\n"
+        "3. 每条信息后附上来源\n"
+        "4. 用'---'分隔不同结果\n"
+        "5. 直接在文本中引用，不使用编号引用\n"
+        "6. 确保提供完整URL"
+    )
+
+    # Use custom system message if provided, otherwise use default
+    system_message = arguments.get("system_message", default_system_message)
+
+    # Prepare full message list with system message first
+    full_messages = [{"role": "system", "content": system_message}]
+    full_messages.extend(user_messages)
+
+    try:
+        # Initialize OpenAI client
+
+        # Make the API call using OpenAI SDK
+        completion = client.chat.completions.create(
+            model="gpt-4o-mini-search-preview",
+            messages=full_messages
+        )
+
+        # Extract content from response
+        content = completion.choices[0].message.content
+
+    except Exception as e:
+        raise RuntimeError(f"API error: {str(e)}")
+
+    return [types.TextContent(
+            type="text",
+            text=content,
+            )]
+
+
+async def main():
+    async with mcp.server.stdio.stdio_server() as (read_stream, write_stream):
+        await server.run(
+            read_stream,
+            write_stream,
+            InitializationOptions(
+                server_name="mcp-server-gpt4o-mini-search",
+                server_version="0.1.0",
+                capabilities=server.get_capabilities(
+                    notification_options=NotificationOptions(
+                        tools_changed=True),
+                    experimental_capabilities={},
+                ),
+            ),
+        )
+
+if __name__ == "__main__":
+    import asyncio
+    asyncio.run(main())

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/rag/api_server.py

@@ -31,6 +31,14 @@ from byzerllm.utils.client.entrypoints.openai.protocol import (
 from pydantic import BaseModel
 from typing import List,Optional
 
+# If support dotenv, use it
+if os.path.exists(".env"):
+    try:
+        from dotenv import load_dotenv
+        load_dotenv()
+    except ImportError:
+        pass
+
 logger = init_logger(__name__)
 
 llm_client: ByzerLLM = None

{auto_coder-0.1.292 → auto_coder-0.1.294}/src/autocoder/rag/cache/local_duckdb_storage_cache.py

@@ -79,7 +79,8 @@ class DuckDBLocalContext:
 class LocalDuckdbStorage:
 
     def __init__(
-            self, llm: Union[ByzerLLM, SimpleByzerLLM] = None, database_name: str = ":memory:", table_name: str = "documents",
+            self, llm: Union[ByzerLLM, SimpleByzerLLM] = None, database_name: str = ":memory:",
+            table_name: str = "documents",
             embed_dim: Optional[int] = None, persist_dir: str = "./storage"
     ) -> None:
         self.llm = llm
@@ -410,9 +411,9 @@ class LocalDuckDBStorageCache(BaseCacheManager):
         from autocoder.rag.token_counter import initialize_tokenizer
 
         with Pool(
-            processes=os.cpu_count(),
-            initializer=initialize_tokenizer,
-            initargs=(VariableHolder.TOKENIZER_PATH,),
+                processes=os.cpu_count(),
+                initializer=initialize_tokenizer,
+                initargs=(VariableHolder.TOKENIZER_PATH,),
         ) as pool:
             target_files_to_process = []
             for file_info in files_to_process:
@@ -451,23 +452,66 @@ class LocalDuckDBStorageCache(BaseCacheManager):
         self.write_cache()
 
         if items:
-            logger.info("Clear cache from Byzer DuckDB Storage")
+            logger.info("[BUILD CACHE] Clearing existing cache from Byzer DuckDB Storage")
             self.storage.truncate_table()
-            logger.info("Save new cache to Byzer DuckDB Storage")
-
-            total_chunks = len(items)
-            completed_chunks = 0
-
-            logger.info(f"进度: 已完成 {0}/{total_chunks} 个文本块")
+            logger.info(f"[BUILD CACHE] Preparing to write to Byzer DuckDB Storage, "
+                        f"total chunks: {len(items)}, total files: {len(files_to_process)}")
+
+            # Use a fixed optimal batch size instead of dividing by worker count
+            batch_size = 100  # Optimal batch size for Byzer Storage
+            item_batches = [items[i:i + batch_size] for i in range(0, len(items), batch_size)]
+
+            total_batches = len(item_batches)
+            completed_batches = 0
+
+            logger.info(f"[BUILD CACHE] Starting to write to Byzer Storage using {batch_size} items per batch, "
+                        f"total batches: {total_batches}")
+            start_time = time.time()
+
+            # Use more workers to process the smaller batches efficiently
+            max_workers = min(10, total_batches)  # Cap at 10 workers or total batch count
+            logger.info(f"[BUILD CACHE] Using {max_workers} parallel workers for processing")
+
+            def batch_add_doc(_batch):
+                for b in _batch:
+                    self.storage.add_doc(b, dim=self.extra_params.rag_duckdb_vector_dim)
+
+            with (ThreadPoolExecutor(max_workers=max_workers) as executor):
+                futures = []
+                # Submit all batches to the executor upfront (non-blocking)
+                for batch in item_batches:
+                    futures.append(
+                        executor.submit(
+                            batch_add_doc, batch
+                        )
+                    )
 
-            for _chunk in items:
-                try:
-                    self.storage.add_doc(_chunk, dim=self.extra_params.rag_duckdb_vector_dim)
-                    completed_chunks += 1
-                    logger.info(f"进度: 已完成 {completed_chunks}/{total_chunks} 个文本块")
-                    time.sleep(self.extra_params.anti_quota_limit)
-                except Exception as err:
-                    logger.error(f"Error in saving chunk: {str(err)}")
+                # Wait for futures to complete
+                for future in as_completed(futures):
+                    try:
+                        future.result()
+                        completed_batches += 1
+                        elapsed = time.time() - start_time
+                        estimated_total = elapsed / completed_batches * total_batches if completed_batches > 0 else 0
+                        remaining = estimated_total - elapsed
+
+                        # Only log progress at reasonable intervals to reduce log spam
+                        if ((completed_batches == 1) or
+                                (completed_batches == total_batches) or
+                                (completed_batches % max(1, total_batches // 10) == 0)):
+                            logger.info(
+                                f"[BUILD CACHE] Progress: {completed_batches}/{total_batches} batches completed "
+                                f"({(completed_batches / total_batches * 100):.1f}%) "
+                                f"Estimated time remaining: {remaining:.1f}s"
+                            )
+                    except Exception as e:
+                        logger.error(f"[BUILD CACHE] Error saving batch: {str(e)}")
+                        # Add more detailed error information
+                        logger.error(f"[BUILD CACHE] Error details: batch size: "
+                                     f"{len(batch) if 'batch' in locals() else 'unknown'}")
+
+            total_time = time.time() - start_time
+            logger.info(f"[BUILD CACHE] All chunks written, total time: {total_time:.2f}s")
 
     def update_storage(self, file_info: FileInfo, is_delete: bool):
         results = self.storage.query_by_path(file_info.file_path)
@@ -596,10 +640,11 @@ class LocalDuckDBStorageCache(BaseCacheManager):
         """Search cached documents using query"""
         self.trigger_update()  # 检查更新
 
-        if options is None or "query" not in options:
+        if options is None or "queries" not in options:
             return {file_path: self.cache[file_path].model_dump() for file_path in self.cache}
 
-        query = options.get("query", "")
+        queries = options.get("queries", "")
+        query = queries[0]
         logger.info(f"正在使用向量搜索检索数据, 你的问题: {query}")
         total_tokens = 0
         results = []
@@ -610,9 +655,9 @@ class LocalDuckDBStorageCache(BaseCacheManager):
             # results = self.storage.vector_search(query, similarity_value=0.7, similarity_top_k=200)
             search_results = self.storage.vector_search(
                 query,
-                similarity_value=self.extra_params.duckdb_query_similarity,
-                similarity_top_k=self.extra_params.duckdb_query_top_k,
-                query_dim=self.extra_params.duckdb_vector_dim
+                similarity_value=self.extra_params.rag_duckdb_query_similarity,
+                similarity_top_k=self.extra_params.rag_duckdb_query_top_k,
+                query_dim=self.extra_params.rag_duckdb_vector_dim
             )
             results.extend(search_results)
 
@@ -644,4 +689,4 @@ class LocalDuckDBStorageCache(BaseCacheManager):
             f"用户Hybrid Index Max Tokens设置为:{self.max_output_tokens}，"
             f"累计tokens: {total_tokens}, "
             f"经过向量搜索共检索出 {len(result.keys())} 个文档, 共 {len(self.cache.keys())} 个文档")
-        return result
+        return result