auto-coder 0.1.260__py3-none-any.whl → 0.1.262__py3-none-any.whl

autocoder/common/auto_coder_lang.py

@@ -3,6 +3,16 @@ from byzerllm.utils import format_str_jinja2
 
 MESSAGES = {
     "en": {
+        "config_validation_error": "Config validation error: {{error}}",
+        "invalid_boolean_value": "Value '{{value}}' is not a valid boolean(true/false)",
+        "invalid_integer_value": "Value '{{value}}' is not a valid integer",
+        "invalid_float_value": "Value '{{value}}' is not a valid float",
+        "invalid_type_value": "Value '{{value}}' is not a valid type (expected: {{types}})",
+        "value_out_of_range": "Value {{value}} is out of allowed range({{min}}~{{max}})",
+        "invalid_choice": "Value '{{value}}' is not in allowed options({{allowed}})",
+        "unknown_config_key": "Unknown config key '{{key}}'",
+        "model_not_found": "Model '{{model}}' is not configured in models.yml",
+        "required_without_default": "Config key '{{key}}' requires explicit value",
         "auto_command_action_break": "Command {{command}} execution failed (got {{action}} result), no result can be obtained, please try again",
         "auto_command_break": "Auto command execution failed to execute command: {{command}}",
         "auto_command_executing": "\n\n============= Executing command: {{command}} =============\n\n",
@@ -123,9 +133,22 @@ MESSAGES = {
         "auto_command_failed": "Auto command failed: {{error}}. Please check your input and try again.",
         "command_execution_result": "{{action}} execution result",
         "satisfied_prompt": "Requirements satisfied, no further action needed",
-        "auto_command_analyzed": "Selected command"
-    },
+        "auto_command_analyzed": "Selected command",
+        "invalid_enum_value": "Value '{{value}}' is not in allowed values ({{allowed}})",
+        "no_changes_made": "⚠️ no changes made, the reason may be that the text block generated by the coding function has a problem, so it cannot be merged into the project",
+        "conversation_pruning_start": "Conversation pruning started, total tokens: {{total_tokens}}, safe zone: {{safe_zone}}",
+        },
     "zh": {
+        "config_validation_error": "配置验证错误: {{error}}",
+        "invalid_boolean_value": "值 '{{value}}' 不是有效的布尔值(true/false)",
+        "invalid_integer_value": "值 '{{value}}' 不是有效的整数",
+        "invalid_float_value": "值 '{{value}}' 不是有效的浮点数",
+        "invalid_type_value": "值 '{{value}}' 不是有效的类型 (期望: {{types}})",
+        "value_out_of_range": "值 {value} 超出允许范围({min}~{max})",
+        "invalid_choice": "值 '{value}' 不在允许选项中({allowed})",
+        "unknown_config_key": "未知的配置项 '{key}'",
+        "model_not_found": "模型 '{model}' 未在 models.yml 中配置",
+        "required_without_default": "配置项 '{key}' 需要明确设置值",
         "auto_command_action_break": "命令 {{command}} 执行失败（获取到了 {{action}} 的结果），无法获得任何结果,请重试",
         "auto_command_break": "自动命令执行失败: {{command}}",
         "auto_command_executing": "\n\n============= 正在执行指令: {{command}} =============\n\n",
@@ -213,7 +236,7 @@ MESSAGES = {
         "unmerged_blocks_warning": "⚠️ 发现 {{ num_blocks }} 个未合并的代码块，更改将不会被应用。请手动检查后重试。",
         "pylint_file_check_failed": "⚠️ {{ file_path }} 的 Pylint 检查失败。更改未应用。错误: {{ error_message }}",
         "merge_success": "✅ 成功合并了 {{ num_files }} 个文件中的更改 {{ num_changes }}/{{ total_blocks }} 个代码块。",
-        "no_changes_made": "⚠️ 未对任何文件进行更改。",
+        "no_changes_made": "⚠️ 未对任何文件进行更改。这个原因可能是因为coding函数生成的文本块格式有问题，导致无法合并进项目",
         "unmerged_blocks_title": "未合并代码块",
         "unmerged_file_path": "文件: {{file_path}}",
         "unmerged_search_block": "Search Block({{similarity}}):",
@@ -245,7 +268,9 @@ MESSAGES = {
         "auto_command_failed": "自动命令执行失败: {{error}}。请检查您的输入并重试。",
         "command_execution_result": "{{action}} 执行结果",
         "satisfied_prompt": "已满足需求，无需进一步操作",
-        "auto_command_analyzed": "被选择指令"
+        "auto_command_analyzed": "被选择指令",
+        "invalid_enum_value": "值 '{{value}}' 不在允许的值列表中 ({{allowed}})",
+        "conversation_pruning_start": "⚠️ 对话长度 {{total_tokens}} tokens 超过安全阈值 {{safe_zone}}，开始修剪对话。"
     }}
 
 

autocoder/common/code_auto_merge.py

@@ -157,7 +157,7 @@ class CodeAutoMerge:
         # get the file name 
         file_name = os.path.basename(self.args.file)
         
-        if not force_skip_git:
+        if not force_skip_git and not self.args.skip_commit:
             try:
                 git_utils.commit_changes(self.args.source_dir, f"auto_coder_pre_{file_name}_{md5}")
             except Exception as e:            
@@ -176,6 +176,6 @@ class CodeAutoMerge:
                 f.write(block.content)
 
         self.printer.print_in_terminal("files_merged", total=total)
-        if not force_skip_git:
+        if not force_skip_git and not self.args.skip_commit:
             commit_result = git_utils.commit_changes(self.args.source_dir, f"auto_coder_{file_name}_{md5}\n{self.args.query}")
             git_utils.print_commit_info(commit_result=commit_result)

autocoder/common/code_auto_merge_diff.py

@@ -511,7 +511,7 @@ class CodeAutoMergeDiff:
         # get the file name 
         file_name = os.path.basename(self.args.file)
         
-        if not force_skip_git:
+        if not force_skip_git and not self.args.skip_commit:
             try:
                 git_utils.commit_changes(self.args.source_dir, f"auto_coder_pre_{file_name}_{md5}")
             except Exception as e:            
@@ -522,6 +522,6 @@ class CodeAutoMergeDiff:
         self.apply_edits(edits)
 
         self.printer.print_in_terminal("files_merged_total", total=total)
-        if not force_skip_git:
+        if not force_skip_git and not self.args.skip_commit:
             commit_result = git_utils.commit_changes(self.args.source_dir, f"auto_coder_{file_name}_{md5}\n{self.args.query}")
             git_utils.print_commit_info(commit_result=commit_result)

autocoder/common/code_auto_merge_editblock.py

@@ -360,7 +360,7 @@ class CodeAutoMergeEditBlock:
                                                   file_path=file_path, 
                                                   error_message=error_message)
 
-        if changes_made and not force_skip_git:
+        if changes_made and not force_skip_git and not self.args.skip_commit:
             try:
                 git_utils.commit_changes(
                     self.args.source_dir, f"auto_coder_pre_{file_name}_{md5}"
@@ -400,7 +400,7 @@ class CodeAutoMergeEditBlock:
             )
 
         if changes_made:
-            if not force_skip_git:
+            if not force_skip_git and not self.args.skip_commit:
                 try:
                     commit_result = git_utils.commit_changes(
                         self.args.source_dir, f"auto_coder_{file_name}_{md5}\n{self.args.query}"

autocoder/common/code_auto_merge_strict_diff.py

@@ -209,7 +209,7 @@ class CodeAutoMergeStrictDiff:
         # get the file name 
         file_name = os.path.basename(self.args.file)
         
-        if not force_skip_git:
+        if not force_skip_git and not self.args.skip_commit:
             try:
                 git_utils.commit_changes(self.args.source_dir, f"auto_coder_pre_{file_name}_{md5}")
             except Exception as e:            
@@ -232,7 +232,7 @@ class CodeAutoMergeStrictDiff:
                 raise Exception("Error applying diff to file: " + path)
                             
         self.printer.print_in_terminal("files_merged_total", total=total)
-        if not force_skip_git:
+        if not force_skip_git and not self.args.skip_commit:
             commit_result = git_utils.commit_changes(self.args.source_dir, f"auto_coder_{file_name}_{md5}\n{self.args.query}")
             git_utils.print_commit_info(commit_result=commit_result)
 

autocoder/common/command_completer.py

@@ -484,16 +484,7 @@ class CommandCompleter(Completer):
                 current_word = parser.current_word()
                 for command in parser.get_sub_commands():
                     if command.startswith(current_word):
-                        yield Completion(command, start_position=-len(current_word))
-
-            elif words[0] == "/coding":
-                new_text = text[len("/coding"):]
-                parser = CommandTextParser(new_text, words[0])
-                parser.lib()
-                current_word = parser.current_word()
-                for command in parser.get_sub_commands():
-                    if command.startswith(current_word):
-                        yield Completion(command, start_position=-len(current_word))
+                        yield Completion(command, start_position=-len(current_word))            
 
             elif words[0] == "/conf":
                 new_words = text[len("/conf"):].strip().split()

autocoder/common/conf_validator.py

@@ -0,0 +1,245 @@
+from typing import Any
+from autocoder.utils import llms as llms_utils
+from autocoder.common.auto_coder_lang import get_message_with_format
+
+class ConfigValidationError(Exception):
+    def __init__(self, message: str):
+        self.message = message
+        super().__init__(message)
+
+class ConfigValidator:
+    CONFIG_SPEC = {
+        # 核心配置项
+        "auto_merge": {
+            "type": str,
+            "allowed": ["editblock", "diff", "wholefile"],
+            "default": "editblock",
+            "description": "代码合并方式(editblock/diff/wholefile)"
+        },
+        "editblock_similarity": {
+            "type": float,
+            "min": 0.0,
+            "max": 1.0,
+            "default": 0.9,
+            "description": "代码块相似度阈值(0-1)"
+        },
+        "generate_times_same_model": {
+            "type": int,
+            "min": 1,
+            "max": 5,
+            "default": 1,
+            "description": "同模型生成次数(1-5)"
+        },
+        "skip_filter_index": {
+            "type": bool,
+            "default": False,
+            "description": "是否跳过根据用户的query自动查找上下文"
+        },
+        "skip_build_index": {
+            "type": bool,
+            "default": True,
+            "description": "是否自动构建索引"
+        },
+        "enable_global_memory": {
+            "type": bool,
+            "default": True,
+            "description": "是否开启全局记忆"
+        },
+        "rank_times_same_model": {
+            "type": int,
+            "min": 1,
+            "max": 3,
+            "default": 1,
+            "description": "相同模型重排序次数"
+        },
+        "human_as_model": {
+            "type": bool,
+            "default": False,
+            "description": "是否以人类作为模型"
+        },
+        "skip_confirm": {
+            "type": bool,
+            "default": True,
+            "description": "是否跳过确认步骤"
+        },
+        "silence": {
+            "type": bool,
+            "default": True,
+            "description": "是否静默模式"
+        },
+        "include_project_structure": {
+            "type": bool,
+            "default": True,
+            "description": "是否包含项目结构"
+        },
+        "product_mode": {
+            "type": str,
+            "allowed": ["lite", "pro"],
+            "default": "lite",
+            "description": "产品模式(lite/pro)"
+        },
+        "model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "默认模型名称"
+        },
+        "chat_model": {
+            "type": str,
+            "default": "r1_chat",
+            "description": "聊天模型名称"
+        },
+        "code_model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "代码生成模型名称"
+        },
+        "index_filter_model": {
+            "type": str,
+            "default": "r1_chat",
+            "description": "索引过滤模型名称"
+        },
+        "generate_rerank_model": {
+            "type": str,
+            "default": "r1_chat",
+            "description": "生成重排序模型名称"
+        },
+        "emb_model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "嵌入模型名称"
+        },
+        "vl_model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "视觉语言模型名称"
+        },
+        "designer_model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "设计模型名称"
+        },
+        "sd_model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "稳定扩散模型名称"
+        },
+        "voice2text_model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "语音转文本模型名称"
+        },
+        "commit_model": {
+            "type": str,
+            "default": "v3_chat",
+            "description": "提交信息生成模型名称"
+        }
+    }
+
+    @classmethod
+    def validate(cls, key: str, value: Any, product_mode: str) -> Any:
+        # 获取配置规范
+        spec = cls.CONFIG_SPEC.get(key)
+        if not spec:
+            # raise ConfigValidationError(
+            #     get_message_with_format("unknown_config_key", key=key)
+            # )
+            return
+
+        # 类型转换和验证
+        try:
+            # 布尔类型特殊处理
+            if isinstance(spec['type'], (list, tuple)):
+                # 多个类型支持
+                for type_ in spec['type']:
+                    try:
+                        if type_ == bool:
+                            return cls.validate_boolean(value)
+                        converted_value = type_(value)
+                        break
+                    except ValueError:
+                        continue
+                else:
+                    types_str = ', '.join([t.__name__ for t in spec['type']])
+                    raise ConfigValidationError(
+                        get_message_with_format(f"invalid_type_value", 
+                                  value=value,
+                                  types=types_str)
+                    )
+            else:
+                # 单个类型处理
+                if spec['type'] == bool:
+                    return cls.validate_boolean(value)
+                converted_value = spec['type'](value)
+        except ValueError:
+            type_name = spec['type'].__name__ if not isinstance(spec['type'], (list, tuple)) else ', '.join([t.__name__ for t in spec['type']])
+            raise ConfigValidationError(
+                get_message_with_format(f"invalid_type_value", 
+                          value=value,
+                          types=type_name)
+            )
+
+        # 范围检查
+        if 'min' in spec and converted_value < spec['min']:
+            raise ConfigValidationError(
+                get_message_with_format("value_out_of_range", 
+                          value=converted_value,
+                          min=spec['min'],
+                          max=spec['max'])
+            )
+        
+        if 'max' in spec and converted_value > spec['max']:
+            raise ConfigValidationError(
+                get_message_with_format("value_out_of_range", 
+                          value=converted_value,
+                          min=spec['min'],
+                          max=spec['max'])
+            )
+
+        # 枚举值检查
+        if 'allowed' in spec and converted_value not in spec['allowed']:
+            raise ConfigValidationError(
+                get_message_with_format("invalid_enum_value", 
+                          value=converted_value,
+                          allowed=', '.join(map(str, spec['allowed'])))
+            )
+        
+        # 模型存在性检查        
+        if product_mode == "lite" and key in ["chat_model","code_model", 
+                                     "index_filter_model", "generate_rerank_model", 
+                                     "rank_times_same_model", 
+                                     "emb_model", "vl_model", "designer_model", "sd_model", 
+                                     "voice2text_model", 
+                                     "commit_model","model"]:
+            if not llms_utils.get_model_info(converted_value,product_mode):
+                raise ConfigValidationError(
+                    get_message_with_format("model_not_found", model=converted_value)
+                )
+
+        return converted_value
+
+    @staticmethod
+    def validate_boolean(value: str) -> bool:
+        if value.lower() in ("true", "1", "yes"):
+            return True
+        if value.lower() in ("false", "0", "no"):
+            return False
+        raise ConfigValidationError(
+            get_message_with_format("invalid_boolean_value", value=value)
+        )
+
+    @classmethod
+    def get_config_docs(cls) -> str:
+        """生成配置项文档"""
+        docs = ["可用配置项："]
+        for key, spec in cls.CONFIG_SPEC.items():
+            desc = [
+                f"- {key}: {spec['description']}",
+                f"  类型: {spec['type'].__name__}",
+                f"  默认值: {spec['default']}"
+            ]
+            if "allowed" in spec:
+                desc.append(f"  允许值: {', '.join(spec['allowed'])}")
+            if "min" in spec and "max" in spec:
+                desc.append(f"  取值范围: {spec['min']}~{spec['max']}")
+            docs.append("\n".join(desc))
+        return "\n\n".join(docs)

autocoder/common/conversation_pruner.py

@@ -0,0 +1,131 @@
+from typing import List, Dict, Any, Union
+import json
+from pydantic import BaseModel
+import byzerllm
+from autocoder.common.printer import Printer
+from autocoder.utils.llms import count_tokens
+from loguru import logger
+
+class PruneStrategy(BaseModel):
+    name: str
+    description: str
+    config: Dict[str, Any] = {"safe_zone_tokens": 0, "group_size": 4}
+
+class ConversationPruner:
+    def __init__(self, llm: Union[byzerllm.ByzerLLM, byzerllm.SimpleByzerLLM], 
+                 safe_zone_tokens: int = 500, group_size: int = 4):
+        self.llm = llm
+        self.printer = Printer()
+        self.strategies = {
+            "summarize": PruneStrategy(
+                name="summarize",
+                description="对早期对话进行分组摘要，保留关键信息",
+                config={"safe_zone_tokens": safe_zone_tokens, "group_size": group_size}
+            ),
+            "truncate": PruneStrategy(
+                name="truncate",
+                description="分组截断最早的部分对话",
+                config={"safe_zone_tokens": safe_zone_tokens, "group_size": group_size}
+            ),
+            "hybrid": PruneStrategy(
+                name="hybrid",
+                description="先尝试分组摘要，如果仍超限则分组截断",
+                config={"safe_zone_tokens": safe_zone_tokens, "group_size": group_size}
+            )
+        }
+
+    def get_available_strategies(self) -> List[Dict[str, Any]]:
+        """获取所有可用策略"""
+        return [strategy.dict() for strategy in self.strategies.values()]
+
+    def prune_conversations(self, conversations: List[Dict[str, Any]], 
+                            strategy_name: str = "summarize") -> List[Dict[str, Any]]:
+        """
+        根据策略修剪对话
+        Args:
+            conversations: 原始对话列表            
+            strategy_name: 策略名称
+        Returns:
+            修剪后的对话列表
+        """        
+        current_tokens = count_tokens(json.dumps(conversations, ensure_ascii=False))
+        if current_tokens <= self.args.conversation_prune_safe_zone_tokens:
+            return conversations
+
+        strategy = self.strategies.get(strategy_name, self.strategies["summarize"])
+        
+        if strategy.name == "summarize":
+            return self._summarize_prune(conversations, strategy.config)
+        elif strategy.name == "truncate":
+            return self._truncate_prune.with_llm(self.llm).run(conversations)
+        elif strategy.name == "hybrid":
+            pruned = self._summarize_prune(conversations, strategy.config)
+            if count_tokens(json.dumps(pruned, ensure_ascii=False)) > self.args.conversation_prune_safe_zone_tokens:
+                return self._truncate_prune(pruned)
+            return pruned
+        else:
+            logger.warning(f"Unknown strategy: {strategy_name}, using summarize instead")
+            return self._summarize_prune(conversations, strategy.config)
+
+    def _summarize_prune(self, conversations: List[Dict[str, Any]], 
+                         config: Dict[str, Any]) -> List[Dict[str, Any]]:
+        """摘要式剪枝"""
+        safe_zone_tokens = config.get("safe_zone_tokens", 50*1024)
+        group_size = config.get("group_size", 4)
+        processed_conversations = conversations.copy()
+        
+        while True:
+            current_tokens = count_tokens(json.dumps(processed_conversations, ensure_ascii=False))
+            if current_tokens <= safe_zone_tokens:
+                break
+                
+            # 找到要处理的对话组
+            early_conversations = processed_conversations[:-group_size]
+            recent_conversations = processed_conversations[-group_size:]
+            
+            if not early_conversations:
+                break
+                
+            # 生成当前组的摘要
+            group_summary = self._generate_summary.with_llm(self.llm).run(early_conversations[-group_size:])
+            
+            # 更新对话历史
+            processed_conversations = early_conversations[:-group_size] + [
+                {"role": "user", "content": f"历史对话摘要：\n{group_summary}"},
+                {"role": "assistant", "content": f"收到"}
+            ] + recent_conversations
+            
+        return processed_conversations
+
+    @byzerllm.prompt()
+    def _generate_summary(self, conversations: List[Dict[str, Any]]) -> str:
+        '''
+        请用中文将以下对话浓缩为要点，保留关键决策和技术细节：
+        
+        <history_conversations>
+        {{conversations}}
+        </history_conversations>
+        '''    
+        return {
+            "conversations": json.dumps(conversations, ensure_ascii=False)
+        }    
+
+    def _truncate_prune(self, conversations: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
+        """截断式剪枝"""
+        safe_zone_tokens = self.strategies["truncate"].config.get("safe_zone_tokens", 0)
+        group_size = self.strategies["truncate"].config.get("group_size", 4)
+        processed_conversations = conversations.copy()
+        
+        while True:
+            current_tokens = count_tokens(json.dumps(processed_conversations, ensure_ascii=False))
+            if current_tokens <= safe_zone_tokens:
+                break
+                
+            # 如果剩余对话不足一组，直接返回
+            if len(processed_conversations) <= group_size:
+                return []
+                
+            # 移除最早的一组对话
+            processed_conversations = processed_conversations[group_size:]
+            
+        return processed_conversations

autocoder/common/index_import_export.py

@@ -0,0 +1,101 @@
+import os
+import json
+import shutil
+from loguru import logger
+from autocoder.common.printer import Printer
+
+
+def export_index(project_root: str, export_path: str) -> bool:
+    printer = Printer()
+    """
+    Export index.json with absolute paths converted to relative paths
+    
+    Args:
+        project_root: Project root directory
+        export_path: Path to export the index file
+        
+    Returns:
+        bool: True if successful, False otherwise
+    """
+    try:
+        index_path = os.path.join(project_root, ".auto-coder", "index.json")
+        if not os.path.exists(index_path):
+            printer.print_in_terminal("index_not_found", path=index_path)
+            return False
+            
+        # Read and convert paths
+        with open(index_path, "r") as f:
+            index_data = json.load(f)
+            
+        # Convert absolute paths to relative
+        converted_data = {}
+        for abs_path, data in index_data.items():
+            try:
+                rel_path = os.path.relpath(abs_path, project_root)
+                data["module_name"] = rel_path
+                converted_data[rel_path] = data
+            except ValueError:
+                printer.print_in_terminal("index_convert_path_fail", path=abs_path)
+                converted_data[abs_path] = data
+        
+        # Write to export location
+        export_file = os.path.join(export_path, "index.json")
+        os.makedirs(export_path, exist_ok=True)
+        with open(export_file, "w") as f:
+            json.dump(converted_data, f, indent=2)
+            
+        return True
+        
+    except Exception as e:
+        printer.print_in_terminal("index_error", error=str(e))
+        return False
+
+def import_index(project_root: str, import_path: str) -> bool:
+    printer = Printer()
+    """
+    Import index.json with relative paths converted to absolute paths
+    
+    Args:
+        project_root: Project root directory
+        import_path: Path containing the index file to import
+        
+    Returns:
+        bool: True if successful, False otherwise
+    """
+    try:
+        import_file = os.path.join(import_path, "index.json")
+        if not os.path.exists(import_file):
+            printer.print_in_terminal("index_not_found", path=import_file)
+            return False
+            
+        # Read and convert paths
+        with open(import_file, "r") as f:
+            index_data = json.load(f)
+            
+        # Convert relative paths to absolute
+        converted_data = {}
+        for rel_path, data in index_data.items():
+            try:
+                abs_path = os.path.join(project_root, rel_path)
+                data["module_name"] = abs_path
+                converted_data[abs_path] = data
+            except Exception:
+                printer.print_in_terminal("index_convert_path_fail", path=rel_path)
+                converted_data[rel_path] = data
+        
+        # Backup existing index
+        index_path = os.path.join(project_root, ".auto-coder", "index.json")
+        if os.path.exists(index_path):
+            backup_path = index_path + ".bak"
+            shutil.copy2(index_path, backup_path)
+            printer.print_in_terminal("index_backup_success", path=backup_path)
+            
+        # Write new index
+        with open(index_path, "w") as f:
+            json.dump(converted_data, f, indent=2)
+            
+        return True
+        
+    except Exception as e:
+        printer.print_in_terminal("index_error", error=str(e))
+        return False