auto-coder 0.1.400__py3-none-any.whl → 2.0.0__py3-none-any.whl

autocoder/rag/types.py

@@ -39,6 +39,22 @@ class OtherStat(BaseModel):
     duration: float = 0.0
 
 
+class ChunkRange(BaseModel):
+    start_line: int
+    end_line: int
+
+
+class ChunkMetadata(BaseModel):
+    original_doc: str
+    chunk_ranges: List[ChunkRange]
+    processing_time: float
+    llm_time: float
+    input_tokens_count: int
+    generated_tokens_count: int
+    duration: float
+    chunk_model: str
+
+
 class RAGStat(BaseModel):
     recall_stat: RecallStat
     chunk_stat: ChunkStat

autocoder/rag/variable_holder.py

@@ -1,3 +1,5 @@
+from typing import Optional
+
 class VariableHolder:
-    TOKENIZER_PATH = None
-    TOKENIZER_MODEL = None
+    TOKENIZER_PATH: Optional[str] = None
+    TOKENIZER_MODEL: Optional[object] = None

autocoder/rags.py

@@ -19,42 +19,44 @@ except ImportError:
     msvcrt = None
 
 
-RAGS_JSON = os.path.expanduser("~/.auto-coder/keys/rags_config.json")
+RAGS_JSON = Path.home() / ".auto-coder" / "keys" / "rags_config.json"
 
 
 class RAGConfigManager:
     """RAG配置管理器，支持并发安全的增删改查操作"""
-    
-    def __init__(self, config_path: str = RAGS_JSON):
+
+    def __init__(self, config_path: str = str(RAGS_JSON)):
         self.config_path = config_path
         self._ensure_config_dir()
-    
+
     def _ensure_config_dir(self):
         """确保配置目录存在"""
-        config_dir = os.path.dirname(self.config_path)
-        if not os.path.exists(config_dir):
-            os.makedirs(config_dir, exist_ok=True)
-    
+        config_path = Path(self.config_path)
+        config_dir = config_path.parent
+        config_dir.mkdir(parents=True, exist_ok=True)
+
     @contextmanager
-    def _file_lock(self, mode='r'):
+    def _file_lock(self, mode="r"):
         """
         文件锁上下文管理器，确保并发安全
         支持Unix (fcntl) 和 Windows (msvcrt) 系统
         """
         # 确保文件存在
-        if not os.path.exists(self.config_path):
-            with open(self.config_path, 'w') as f:
+        config_path = Path(self.config_path)
+        if not config_path.exists():
+            config_path.parent.mkdir(parents=True, exist_ok=True)
+            with open(config_path, "w") as f:
                 json.dump({}, f)
-        
-        file_handle = open(self.config_path, mode)
+
+        file_handle = open(config_path, mode)
         lock_acquired = False
-        
+
         try:
             # 尝试获取文件锁
             if platform.system() != "Windows":
                 # Unix系统使用fcntl
                 if fcntl is not None:
-                    if 'w' in mode or 'a' in mode or '+' in mode:
+                    if "w" in mode or "a" in mode or "+" in mode:
                         fcntl.flock(file_handle.fileno(), fcntl.LOCK_EX)  # 独占锁
                     else:
                         fcntl.flock(file_handle.fileno(), fcntl.LOCK_SH)  # 共享锁
@@ -83,7 +85,9 @@ class RAGConfigManager:
                                     # 最后一次尝试使用阻塞锁
                                     try:
                                         file_handle.seek(0)
-                                        msvcrt.locking(file_handle.fileno(), msvcrt.LK_LOCK, 1)
+                                        msvcrt.locking(
+                                            file_handle.fileno(), msvcrt.LK_LOCK, 1
+                                        )
                                         lock_acquired = True
                                     except IOError:
                                         # 如果仍然失败，继续执行但不使用锁
@@ -91,11 +95,11 @@ class RAGConfigManager:
                     except Exception:
                         # 如果出现任何异常，继续执行但不使用锁
                         pass
-            
+
             # 重置文件位置到开始
             file_handle.seek(0)
             yield file_handle
-            
+
         finally:
             try:
                 # 释放文件锁
@@ -116,205 +120,210 @@ class RAGConfigManager:
                 pass
             finally:
                 file_handle.close()
-    
+
     def _load_config(self) -> Dict[str, Any]:
         """
         加载配置文件
         """
         try:
-            with self._file_lock('r') as f:
+            with self._file_lock("r") as f:
                 content = f.read().strip()
                 if not content:
                     return {}
                 return json.loads(content)
         except (FileNotFoundError, json.JSONDecodeError):
             return {}
-    
+
     def _save_config(self, config: Dict[str, Any]):
         """
         保存配置文件
         """
-        with self._file_lock('w') as f:
+        with self._file_lock("w") as f:
             json.dump(config, f, indent=2, ensure_ascii=False)
-    
+
     def create(self, name: str, config: Dict[str, Any]) -> bool:
         """
         创建新的RAG服务配置
-        
+
         Args:
             name: RAG服务名称
             config: 配置字典
-            
+
         Returns:
             bool: 创建成功返回True，如果名称已存在返回False
         """
         all_configs = self._load_config()
-        
+
         if name in all_configs:
             return False
-        
+
         # 确保配置包含必要的字段
-        config['name'] = name
-        if 'status' not in config:
-            config['status'] = 'stopped'
-        if 'created_at' not in config:
-            config['created_at'] = datetime.now().isoformat()
-        config['updated_at'] = datetime.now().isoformat()
-        
+        config["name"] = name
+        if "status" not in config:
+            config["status"] = "stopped"
+        if "created_at" not in config:
+            config["created_at"] = datetime.now().isoformat()
+        config["updated_at"] = datetime.now().isoformat()
+
         all_configs[name] = config
         self._save_config(all_configs)
         return True
-    
+
     def read(self, name: Optional[str] = None) -> Optional[Dict[str, Any]]:
         """
         读取RAG服务配置
-        
+
         Args:
             name: RAG服务名称，如果为None则返回所有配置
-            
+
         Returns:
             Dict: 配置字典，如果name不存在或为None且无配置则返回None
         """
         all_configs = self._load_config()
-        
+
         if name is None:
             return all_configs if all_configs else None
-        
+
         return all_configs.get(name)
-    
+
     def update(self, name: str, config: Dict[str, Any]) -> bool:
         """
         更新RAG服务配置
-        
+
         Args:
             name: RAG服务名称
             config: 新的配置字典（会与现有配置合并）
-            
+
         Returns:
             bool: 更新成功返回True，如果名称不存在返回False
         """
         all_configs = self._load_config()
-        
+
         if name not in all_configs:
             return False
-        
+
         # 合并配置
         existing_config = all_configs[name]
         existing_config.update(config)
-        existing_config['name'] = name  # 确保名称不被覆盖
-        existing_config['updated_at'] = datetime.now().isoformat()
-        
+        existing_config["name"] = name  # 确保名称不被覆盖
+        existing_config["updated_at"] = datetime.now().isoformat()
+
         all_configs[name] = existing_config
         self._save_config(all_configs)
         return True
-    
+
     def delete(self, name: str) -> bool:
         """
         删除RAG服务配置
-        
+
         Args:
             name: RAG服务名称
-            
+
         Returns:
             bool: 删除成功返回True，如果名称不存在返回False
         """
         all_configs = self._load_config()
-        
+
         if name not in all_configs:
             return False
-        
+
         del all_configs[name]
         self._save_config(all_configs)
         return True
-    
+
     def list_names(self) -> List[str]:
         """
         获取所有RAG服务名称列表
-        
+
         Returns:
             List[str]: 服务名称列表
         """
         all_configs = self._load_config()
         return list(all_configs.keys())
-    
+
     def exists(self, name: str) -> bool:
         """
         检查RAG服务是否存在
-        
+
         Args:
             name: RAG服务名称
-            
+
         Returns:
             bool: 存在返回True，否则返回False
         """
         all_configs = self._load_config()
         return name in all_configs
-    
+
     def get_by_port(self, port: int) -> Optional[Dict[str, Any]]:
         """
         根据端口号查找RAG服务配置
-        
+
         Args:
             port: 端口号
-            
+
         Returns:
             Dict: 配置字典，如果未找到返回None
         """
         all_configs = self._load_config()
-        
+
         for name, config in all_configs.items():
-            if config.get('port') == port:
+            if config.get("port") == port:
                 return config
-        
+
         return None
-    
+
     def get_running_services(self) -> Dict[str, Any]:
         """
         获取所有运行中的RAG服务
-        
+
         Returns:
             Dict: 运行中的服务配置字典
         """
         all_configs = self._load_config()
         running_services = {}
-        
+
         for name, config in all_configs.items():
-            if config.get('status') == 'running':
+            if config.get("status") == "running":
                 running_services[name] = config
-        
+
         return running_services
-    
+
     def update_status(self, name: str, status: str, **kwargs) -> bool:
         """
         更新RAG服务状态
-        
+
         Args:
             name: RAG服务名称
             status: 新状态 ('running', 'stopped', 'error', etc.)
             **kwargs: 其他要更新的状态信息（如process_id, stdout_fd等）
-            
+
         Returns:
             bool: 更新成功返回True，如果名称不存在返回False
         """
-        update_data = {'status': status, **kwargs}
+        update_data = {"status": status, **kwargs}
         return self.update(name, update_data)
-    
+
     def cleanup_stopped_services(self):
         """
         清理已停止服务的临时状态信息（如process_id, stdout_fd等）
         """
         all_configs = self._load_config()
         modified = False
-        
+
         for name, config in all_configs.items():
-            if config.get('status') == 'stopped':
+            if config.get("status") == "stopped":
                 # 清理临时状态字段
-                temp_fields = ['process_id', 'stdout_fd', 'stderr_fd', 'cache_build_task_id']
+                temp_fields = [
+                    "process_id",
+                    "stdout_fd",
+                    "stderr_fd",
+                    "cache_build_task_id",
+                ]
                 for field in temp_fields:
                     if field in config:
                         del config[field]
                         modified = True
-        
+
         if modified:
             self._save_config(all_configs)
 
@@ -367,5 +376,3 @@ def get_running_rags() -> Dict[str, Any]:
 def update_rag_status(name: str, status: str, **kwargs) -> bool:
     """更新RAG服务状态的便捷函数"""
     return rag_manager.update_status(name, status, **kwargs)
-
-

autocoder/regexproject/__init__.py

@@ -10,10 +10,9 @@ from autocoder.common.search import Search, SearchEngine
 from loguru import logger
 from pydantic import BaseModel, Field
 from rich.console import Console
-import json
-from autocoder.utils.queue_communicate import queue_communicate, CommunicateEvent, CommunicateEventType
 from autocoder.common import files as FileUtils
 
+
 class RegPattern(BaseModel):
     pattern: str = Field(
         ...,
@@ -78,7 +77,7 @@ class RegexProject:
             raise ValueError("Invalid project_type format. Expected 'regex//<pattern>'")
 
     def output(self):
-        with open(self.target_file, "r",encoding="utf-8") as file:
+        with open(self.target_file, "r", encoding="utf-8") as file:
             return file.read()
 
     def is_regex_match(self, file_path):
@@ -93,7 +92,7 @@ class RegexProject:
         return SourceCode(module_name=module_name, source_code=source_code)
 
     def get_source_codes(self) -> Generator[SourceCode, None, None]:
-        for root, dirs, files in os.walk(self.directory,followlinks=True):
+        for root, dirs, files in os.walk(self.directory, followlinks=True):
             for file in files:
                 file_path = os.path.join(root, file)
                 if self.is_regex_match(file_path):
@@ -120,19 +119,22 @@ class RegexProject:
     def get_rag_source_codes(self):
         if not self.args.enable_rag_search and not self.args.enable_rag_context:
             return []
-            
+
         console = Console()
-        console.print(f"\n[bold blue]Starting RAG search for:[/bold blue] {self.args.query}")
-            
+        console.print(
+            f"\n[bold blue]Starting RAG search for:[/bold blue] {self.args.query}"
+        )
+
         from autocoder.rag.rag_entry import RAGFactory
+
         rag = RAGFactory.get_rag(self.llm, self.args, "")
         docs = rag.search(self.args.query)
         for doc in docs:
             doc.tag = "RAG"
-            
+
         console = Console()
         console.print(f"[bold green]Found {len(docs)} relevant documents[/bold green]")
-            
+
         return docs
 
     def get_search_source_codes(self):
@@ -213,20 +215,20 @@ class RegexProject:
         if self.git_url is not None:
             self.clone_repository()
 
-        if self.target_file:
-            with open(self.target_file, "w",encoding="utf-8") as file:
-                for code in self.get_source_codes():
-                    self.sources.append(code)
-                    file.write(f"##File: {code.module_name}\n")
-                    file.write(f"{code.source_code}\n\n")
+        # 先收集所有sources
+        for code in self.get_source_codes():
+            self.sources.append(code)
 
-                for code in self.get_rest_source_codes():
-                    self.sources.append(code)
-                    file.write(f"##File: {code.module_name}\n")
-                    file.write(f"{code.source_code}\n\n")
+        for code in self.get_rest_source_codes():
+            self.sources.append(code)
+
+        for code in self.get_search_source_codes():
+            self.sources.append(code)
 
-                for code in self.get_search_source_codes():
-                    self.sources.append(code)
+        # 如果target_file存在，才写入文件
+        if self.target_file:
+            with open(self.target_file, "w", encoding="utf-8") as file:
+                for code in self.sources:
                     file.write(f"##File: {code.module_name}\n")
                     file.write(f"{code.source_code}\n\n")
 

autocoder/run_context.py

@@ -10,6 +10,7 @@ class RunMode(Enum):
     TERMINAL = auto()
     WEB = auto()
     API = auto()
+    CLI = auto()
 
 
 class RunContext:
@@ -52,6 +53,14 @@ class RunContext:
         """Check if running in web mode."""
         return self._mode == RunMode.WEB
 
+    def is_cli(self) -> bool:
+        """Check if running in cli mode."""
+        return self._mode == RunMode.CLI
+        
+    def is_api(self) -> bool:
+        """Check if running in api mode."""
+        return self._mode == RunMode.API        
+
 
 def get_run_context() -> RunContext:
     """