auto-coder-web 0.1.95__py3-none-any.whl → 0.1.97__py3-none-any.whl

auto_coder_web/agentic_message_parser/__init__.py

@@ -0,0 +1,10 @@
+"""
+Message parser module for processing agentic messages.
+This module provides functionality to parse and process messages from different tools.
+"""
+from .message_parser import parse_message, register_parser, parse_messages
+
+# Import tool parsers to register them
+from . import tool_parsers
+
+__all__ = ['parse_message', 'register_parser', 'parse_messages']

auto_coder_web/agentic_message_parser/message_parser.py

@@ -0,0 +1,79 @@
+"""
+Message parser for processing agentic messages.
+
+This module provides a registry-based approach to parse and process messages
+from different tools. New parsers can be easily registered to handle different
+tool types.
+"""
+import json
+from typing import Dict, Any, Callable, Optional, List, TypeVar, cast
+
+# Define a type for parser functions
+T = TypeVar('T')
+ParserFunc = Callable[[Dict[str, Any], Dict[str, Any]], Optional[Dict[str, Any]]]
+
+# Registry to store message parsers
+_PARSERS: Dict[str, ParserFunc] = {}
+
+def register_parser(tool_name: str):
+    """
+    Decorator to register a parser function for a specific tool.
+    
+    Args:
+        tool_name: The name of the tool this parser handles
+        
+    Returns:
+        Decorator function
+    """
+    def decorator(func: ParserFunc) -> ParserFunc:
+        _PARSERS[tool_name] = func
+        return func
+    return decorator
+
+def parse_message(message: Dict[str, Any]) -> Dict[str, Any]:
+    """
+    Parse a message and apply the appropriate parser based on the tool_name.
+    
+    Args:
+        message: The message to parse
+        
+    Returns:
+        The processed message
+    """
+    processed_message = message.copy()
+    
+    try:
+        # Try to parse the message content as JSON
+        content = message.get("content", "")
+        if not isinstance(content, str):
+            return processed_message
+            
+        content_obj = json.loads(content)
+        
+        # Try all registered parsers
+        for tool_name, parser in _PARSERS.items():
+            # Let each parser decide if it can handle this message
+            result = parser(content_obj, message)
+            if result is not None:
+                return result
+    
+    except (json.JSONDecodeError, TypeError, AttributeError):
+        # If parsing fails, keep the original message unchanged
+        pass
+        
+    return processed_message
+
+def parse_messages(messages: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
+    """
+    Parse a list of messages, applying the appropriate parser to each.
+    
+    Args:
+        messages: List of messages to parse
+        
+    Returns:
+        List of processed messages
+    """
+    return [parse_message(message) for message in messages]
+
+# Tool-specific parsers are defined in tool_parsers.py
+# and automatically registered when that module is imported

auto_coder_web/agentic_message_parser/tool_parsers.py

@@ -0,0 +1,70 @@
+"""
+Tool-specific parsers for processing messages from different tools.
+
+This module contains parser implementations for various tools.
+New parsers can be added here and will be automatically registered.
+"""
+import json
+from typing import Dict, Any, Optional
+from .message_parser import register_parser
+
+@register_parser("ReadFileTool")
+def read_file_tool_parser(content_obj: Dict[str, Any], message: Dict[str, Any]) -> Optional[Dict[str, Any]]:
+    """
+    Parser for ReadFileTool messages.
+    Truncates file content to 200 characters if it's too long.
+    
+    Args:
+        content_obj: The parsed content object
+        message: The original message
+        
+    Returns:
+        The processed message if this parser can handle it, None otherwise
+    """
+    # Validate if this is a ReadFileTool message
+    if not (isinstance(content_obj, dict) and
+            content_obj.get("tool_name") == "ReadFileTool" and
+            "success" in content_obj and
+            "message" in content_obj and
+            "content" in content_obj):
+        return None
+    
+    # Process the content
+    processed_message = message.copy()
+    if isinstance(content_obj["content"], str) and len(content_obj["content"]) > 200:
+        content_obj["content"] = content_obj["content"][:200] + "..."
+        processed_message["content"] = json.dumps(content_obj)
+    
+    return processed_message
+
+# Example of how to add more parsers in the future:
+# 
+# @register_parser("CodeSearchTool")
+# def code_search_tool_parser(content_obj: Dict[str, Any], message: Dict[str, Any]) -> Optional[Dict[str, Any]]:
+#     """
+#     Parser for CodeSearchTool messages.
+#     Truncates search results if they're too long.
+#     
+#     Args:
+#         content_obj: The parsed content object
+#         message: The original message
+#         
+#     Returns:
+#         The processed message if this parser can handle it, None otherwise
+#     """
+#     # Validate if this is a CodeSearchTool message
+#     if not (isinstance(content_obj, dict) and
+#             content_obj.get("tool_name") == "CodeSearchTool" and
+#             "success" in content_obj and
+#             "message" in content_obj and
+#             "content" in content_obj):
+#         return None
+#     
+#     # Process the content
+#     processed_message = message.copy()
+#     if isinstance(content_obj["content"], list) and len(content_obj["content"]) > 5:
+#         content_obj["content"] = content_obj["content"][:5]
+#         content_obj["message"] = f"Showing first 5 of {len(content_obj['content'])} results"
+#         processed_message["content"] = json.dumps(content_obj)
+#     
+#     return processed_message

auto_coder_web/common_router/chat_list_manager.py

@@ -16,7 +16,7 @@ def _get_chat_list_file_path(project_path: str, name: str) -> str:
     chat_lists_dir = _get_chat_lists_dir(project_path)
     return os.path.join(chat_lists_dir, f"{name}.json")
 
-async def save_chat_list(project_path: str, name: str, messages: List[Dict[str, Any]]) -> None:
+async def save_chat_list(project_path: str, name: str, messages: List[Dict[str, Any]], metadata: dict = None) -> None:
     """
     保存聊天列表到文件
     
@@ -24,14 +24,21 @@ async def save_chat_list(project_path: str, name: str, messages: List[Dict[str,
         project_path: 项目路径
         name: 聊天列表名称
         messages: 聊天消息列表
+        metadata: 聊天元数据
         
     Raises:
         Exception: 如果保存失败
     """
     file_path = _get_chat_list_file_path(project_path, name)
     try:
+        data = {
+            "name": name,
+            "messages": messages
+        }
+        if metadata is not None:
+            data["metadata"] = metadata
         async with aiofiles.open(file_path, 'w') as f:
-            await f.write(json.dumps({"messages": messages}, indent=2, ensure_ascii=False))
+            await f.write(json.dumps(data, indent=2, ensure_ascii=False))
     except Exception as e:
         logger.error(f"Error saving chat list {name}: {str(e)}")
         raise e
@@ -73,18 +80,7 @@ async def get_chat_lists(project_path: str) -> List[str]:
 
 async def get_chat_list(project_path: str, name: str) -> Dict[str, Any]:
     """
-    获取特定聊天列表的内容
-    
-    Args:
-        project_path: 项目路径
-        name: 聊天列表名称
-        
-    Returns:
-        聊天列表内容
-        
-    Raises:
-        FileNotFoundError: 如果聊天列表不存在
-        Exception: 如果读取失败
+    获取特定聊天列表的内容（兼容旧结构）
     """
     file_path = _get_chat_list_file_path(project_path, name)
     if not os.path.exists(file_path):
@@ -93,7 +89,13 @@ async def get_chat_list(project_path: str, name: str) -> Dict[str, Any]:
     try:
         async with aiofiles.open(file_path, 'r') as f:
             content = await f.read()
-            return json.loads(content)
+            data = json.loads(content)
+            # 兼容旧数据结构（只有messages）
+            if "name" not in data:
+                data["name"] = name
+            if "metadata" not in data:
+                data["metadata"] = None
+            return data
     except json.JSONDecodeError as e:
         logger.error(f"Invalid JSON in chat list {name}: {str(e)}")
         raise Exception(f"Invalid JSON in chat list file: {str(e)}")

auto_coder_web/common_router/chat_list_router.py

@@ -33,8 +33,8 @@ router = APIRouter()
 @router.post("/api/chat-lists/save")
 async def save_chat_list_endpoint(chat_list: ChatList, project_path: str = Depends(get_project_path)):
     try:
-        # 调用管理模块保存聊天列表
-        await save_chat_list(project_path, chat_list.name, chat_list.messages)
+        # 调用管理模块保存聊天列表，支持 metadata
+        await save_chat_list(project_path, chat_list.name, chat_list.messages, metadata=chat_list.metadata.dict() if chat_list.metadata else None)
         return {"status": "success", "message": f"Chat list {chat_list.name} saved successfully"}
     except Exception as e:
         raise HTTPException(status_code=500, detail=str(e))

auto_coder_web/common_router/file_group_router.py

@@ -8,6 +8,11 @@ import os
 from autocoder.rag.token_counter import count_tokens
 import aiofiles
 from loguru import logger
+from autocoder.rag.loaders import (
+    extract_text_from_pdf,
+    extract_text_from_docx,
+    extract_text_from_ppt
+)
 
 router = APIRouter()
 
@@ -147,6 +152,53 @@ async def auto_create_groups(
         raise HTTPException(status_code=500, detail=str(e))
 
 
+async def _read_file(file_path_to_read: str) -> str:
+    """使用线程模拟异步读取文件内容
+    
+    Args:
+        file_path_to_read: 要读取的文件路径
+        
+    Returns:
+        str: 文件内容
+    """
+    ext = os.path.splitext(file_path_to_read)[1].lower()
+    
+    # 定义各种文件类型的读取函数
+    def read_pdf():
+        logger.info(f"Extracting text from PDF: {file_path_to_read}")
+        return extract_text_from_pdf(file_path_to_read)
+        
+    def read_docx():
+        logger.info(f"Extracting text from DOCX: {file_path_to_read}")
+        return extract_text_from_docx(file_path_to_read)
+        
+    def read_ppt():
+        logger.info(f"Extracting text from PPT/PPTX: {file_path_to_read}")
+        slide_texts = []
+        for slide_identifier, slide_text_content in extract_text_from_ppt(file_path_to_read):
+            slide_texts.append(f"--- Slide {slide_identifier} ---\n{slide_text_content}")
+        return "\n\n".join(slide_texts) if slide_texts else ""
+        
+    def read_text():
+        logger.info(f"Reading plain text file: {file_path_to_read}")
+        with open(file_path_to_read, 'r', encoding='utf-8', errors='replace') as f:
+            return f.read()
+    
+    # 根据文件类型选择相应的读取函数
+    if ext == '.pdf':
+        read_func = read_pdf
+    elif ext == '.docx':
+        read_func = read_docx
+    elif ext in ('.pptx', '.ppt'):
+        read_func = read_ppt
+    else:
+        read_func = read_text
+    
+    # 使用线程执行耗时的文件读取操作
+    content = await asyncio.to_thread(read_func)
+    return content
+
+
 async def count_tokens_from_file(file_path: str) -> int:
     """异步计算文件的token数
     
@@ -156,19 +208,17 @@ async def count_tokens_from_file(file_path: str) -> int:
     Returns:
         int: token数量，出错时返回0
     """
-    try:
-        if not os.path.exists(file_path):
-            logger.warning(f"文件不存在: {file_path}")
-            return 0
-            
+    try:            
         logger.info(f"计算文件token: {file_path}")
-        async with aiofiles.open(file_path, 'r', encoding='utf-8') as f:
-            content = await f.read()
+        content = await _read_file(file_path)
+        
+        if content is None:
+            return 0
         
         file_tokens = count_tokens(content)
         return file_tokens if file_tokens > 0 else 0
     except Exception as e:
-        logger.error(f"读取或计算文件token出错: {file_path}, 错误: {str(e)}")
+        logger.error(f"计算文件token出错: {file_path}, 错误: {str(e)}")
         return 0
 
 

auto_coder_web/routers/auto_router.py

@@ -69,22 +69,30 @@ def ensure_task_dir(project_path: str) -> str:
 
 @byzerllm.prompt()
 def coding_prompt(messages: List[Dict[str, Any]], query: str):
-    '''
-    下面是我们已经产生的一个消息列表,其中 USER_RESPONSE 表示用户的输入，其他都是你的输出：
-    <messages>
+    '''        
+    【历史对话】按时间顺序排列，从旧到新：
     {% for message in messages %}
     <message>
-    <type>{{ message.type }}</type>
-    <content>{{ message.content }}</content>
+    {% if message.type == "USER" or message.type == "USER_RESPONSE" or message.metadata.path == "/agent/edit/tool/result" %}【用户】{% else %}【助手】{% endif %}    
+    <content>
+    {{ message.content }}
+    </content>
     </message>
     {% endfor %}
-    </messages>
     
-    下面是用户的最新需求：
-    <request>
-    {{ query }}    
-    </request>
+    【当前问题】用户的最新需求如下:
+    <current_query>
+    {{ query }}
+    </current_query>            
     '''
+    # 使用消息解析器处理消息
+    from auto_coder_web.agentic_message_parser import parse_messages
+    processed_messages = parse_messages(messages)
+    
+    return {
+        "messages": processed_messages,
+        "query": query
+    }
 
 
 
@@ -148,13 +156,13 @@ async def auto_command(request: AutoCommandRequest, project_path: str = Depends(
 
             # 调用auto_command_wrapper方法  
             logger.info(f"Executing auto command {file_id} with prompt: {prompt_text}")          
-            result = wrapper.auto_command_wrapper(prompt_text, {
+            wrapper.auto_command_wrapper(prompt_text, {
                 "event_file_id": file_id
             })            
-            get_event_manager(event_file).write_completion(
-                EventContentCreator.create_completion(
-                    "200", "completed", result).to_dict()
-            )
+            # get_event_manager(event_file).write_completion(
+            #     EventContentCreator.create_completion(
+            #         "200", "completed", result).to_dict()
+            # )
             logger.info(f"Event file id: {file_id} completed successfully")
         except Exception as e:
             logger.error(f"Error executing auto command {file_id}: {str(e)}")

auto_coder_web/types.py

@@ -24,9 +24,15 @@ class CompletionResponse(BaseModel):
     completions: List[CompletionItem]
 
 
+class ChatMetadata(BaseModel):
+    token_usage: Optional[int] = None  # token 消耗
+    cost: Optional[float] = None       # 费用
+    window_size: Optional[int] = None  # 窗口大小
+
 class ChatList(BaseModel):
     name: str
     messages: List[Dict[str, Any]]
+    metadata: Optional[ChatMetadata] = None  # 新增 metadata 字段
 
 
 class HistoryQuery(BaseModel):

auto_coder_web/version.py

@@ -1 +1 @@
-__version__ = "0.1.95"
+__version__ = "0.1.97"