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

autocoder/common/pruner/tools/test_message_ids_pruning_only.py

@@ -0,0 +1,192 @@
+#!/usr/bin/env python3
+"""
+专门测试消息ID裁剪逻辑的验证脚本
+
+只测试 _apply_message_ids_pruning 方法，不受 token 阈值限制影响
+
+支持的消息ID格式：
+1. 直接字段格式: {"message_id": "9226b3a4-1234-5678-9abc-def012345678"}
+2. Hint嵌入格式: {"content": "消息内容 [[message_id: 9226b3a4]]"}
+3. 兼容格式: {"content": "消息内容 message_id: 9226b3a4"}
+"""
+
+import sys
+import os
+from datetime import datetime
+
+# 添加src路径到Python路径
+sys.path.insert(0, os.path.join(os.path.dirname(os.path.abspath(__file__)), '../../../..'))
+
+from autocoder.common.pruner.agentic_conversation_pruner import AgenticConversationPruner
+from autocoder.common.pruner.conversation_message_ids_api import get_conversation_message_ids_api
+from autocoder.common import AutoCoderArgs
+from autocoder.sdk import get_llm
+
+
+def test_message_ids_pruning_logic():
+    """直接测试消息ID裁剪逻辑"""
+    
+    print("🧪 消息ID裁剪逻辑专项测试")
+    print(f"📅 测试时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
+    print("=" * 80)
+    
+    # 使用已知的会话ID
+    conversation_id = "796e1c2e-1ab8-46d1-9448-f9bd29d5c095"
+    
+    # 创建测试数据，包含要删除的消息ID
+    # 展示两种消息ID格式的支持
+    test_conversations = [
+        {
+            "role": "system",
+            "content": "You are a helpful assistant.",
+            "message_id": "12345678abcdef01"  # 格式1: 直接字段格式（完整UUID）
+        },
+        {
+            "role": "user", 
+            "content": "Hello, can you help me with some code? [[message_id: 87654321]]",  # 格式2: 标准hint格式
+            "message_id": "87654321fedcba02"  # 同时测试优先级（字段优先）
+        },
+        {
+            "role": "assistant",
+            "content": "Of course! I'd be happy to help you with your code. What do you need assistance with?",
+            "message_id": "642f04ee11223344"  # 格式1: 这个消息ID应该被删除
+        },
+        {
+            "role": "user",
+            "content": "I want to read a file and analyze it. message_id: 13579bdf",  # 格式2: 兼容hint格式
+            # 没有message_id字段，只从content中提取
+        },
+        {
+            "role": "assistant", 
+            "content": "I'll help you read and analyze the file. [[message_id: 97531eca]]",  # 格式2: 标准hint格式
+            # 没有message_id字段，只从content中提取
+        }
+    ]
+    
+    print("📋 测试数据:")
+    for i, conv in enumerate(test_conversations, 1):
+        role = conv.get("role", "unknown")
+        message_id = conv.get("message_id", "no_id")[:8]
+        content = conv.get("content", "")[:50] + "..." if len(conv.get("content", "")) > 50 else conv.get("content", "")
+        print(f"  {i}. [{role:9}] {message_id} | {content}")
+    
+    # 查询消息ID配置
+    api = get_conversation_message_ids_api()
+    message_ids_config = api.get_conversation_message_ids(conversation_id)
+    
+    if not message_ids_config:
+        print(f"❌ 未找到会话ID {conversation_id} 的消息删除配置")
+        return False
+    
+    print(f"\n✅ 消息ID配置:")
+    print(f"   要删除的消息ID: {message_ids_config.message_ids}")
+    print(f"   保持成对删除: {message_ids_config.preserve_pairs}")
+    
+    # 创建裁剪器
+    args = AutoCoderArgs(
+        source_dir=".",
+        conversation_prune_safe_zone_tokens=1000,
+        context_prune=True,
+        context_prune_strategy="extract",
+        query="测试消息ID裁剪逻辑"
+    )
+    llm = get_llm("v3_chat", product_mode="lite")
+    
+    # 创建裁剪器
+    from autocoder.common.conversations.get_conversation_manager import get_conversation_manager
+    
+    # 设置当前对话ID
+    conversation_manager = get_conversation_manager()
+    conversation_manager.set_current_conversation(conversation_id)
+    
+    # 创建AgenticConversationPruner实例并执行测试
+    pruner = AgenticConversationPruner(args=args, llm=llm, conversation_id=conversation_id)
+    
+    # 直接测试消息ID裁剪逻辑
+    print(f"\n🚀 执行消息ID裁剪:")
+    
+    try:
+        # 由于 _apply_message_ids_pruning 是私有方法，我们通过完整流程测试但强制低token阈值
+        # 设置一个极小的token阈值来强制触发裁剪逻辑
+        args.conversation_prune_safe_zone_tokens = 1  # 极小值，确保触发
+        
+        # 使用完整的 prune_conversations 方法
+        pruned_conversations = pruner.prune_conversations(test_conversations)
+        
+        print(f"\n📊 裁剪结果:")
+        print(f"   原始消息数量: {len(test_conversations)}")
+        print(f"   裁剪后数量: {len(pruned_conversations)}")
+        print(f"   删除的消息数: {len(test_conversations) - len(pruned_conversations)}")
+        
+        # 检查具体被删除的消息
+        original_ids = [conv.get("message_id", "")[:8] for conv in test_conversations]
+        pruned_ids = [conv.get("message_id", "")[:8] for conv in pruned_conversations]
+        deleted_ids = [msg_id for msg_id in original_ids if msg_id not in pruned_ids]
+        remaining_ids = pruned_ids
+        
+        print(f"\n🔍 详细分析:")
+        print(f"   原始消息ID: {original_ids}")
+        print(f"   删除的消息ID: {deleted_ids}")
+        print(f"   保留的消息ID: {remaining_ids}")
+        
+        # 验证目标消息ID是否被删除
+        target_id = "642f04ee"
+        if target_id in deleted_ids:
+            print(f"   ✅ 目标消息ID {target_id} 已成功删除")
+        else:
+            print(f"   ❌ 目标消息ID {target_id} 未被删除")
+            return False
+        
+        # 验证成对删除逻辑
+        if message_ids_config.preserve_pairs:
+            print(f"\n🔗 成对删除验证:")
+            if "87654321" in deleted_ids and "642f04ee" in deleted_ids:
+                print(f"   ✅ 成对删除正确：user消息 87654321 和 assistant消息 642f04ee 都被删除")
+            else:
+                print(f"   ⚠️ 成对删除检查：删除的消息ID = {deleted_ids}")
+        
+        # 显示裁剪后的消息列表
+        print(f"\n📋 裁剪后的消息:")
+        for i, conv in enumerate(pruned_conversations, 1):
+            role = conv.get("role", "unknown")
+            message_id = conv.get("message_id", "no_id")[:8]
+            content = conv.get("content", "")[:50] + "..." if len(conv.get("content", "")) > 50 else conv.get("content", "")
+            print(f"  {i}. [{role:9}] {message_id} | {content}")
+        
+        # 获取裁剪统计信息
+        stats = pruner.get_pruning_statistics()
+        if stats["range_pruning"]["applied"]:
+            print(f"\n📈 裁剪统计:")
+            print(f"   消息ID裁剪已应用: {stats['range_pruning']['applied']}")
+            print(f"   消息ID裁剪成功: {stats['range_pruning']['success']}")
+            print(f"   总压缩比: {stats['compression']['total_compression_ratio']:.2%}")
+        
+        return True
+        
+    except Exception as e:
+        print(f"❌ 消息ID裁剪测试失败: {str(e)}")
+        import traceback
+        traceback.print_exc()
+        return False
+
+
+def main():
+    """主函数"""
+    success = test_message_ids_pruning_logic()
+    
+    print(f"\n🎯 测试结果: {'✅ 通过' if success else '❌ 失败'}")
+    
+    if success:
+        print("\n🎉 消息ID裁剪逻辑工作正常！")
+        print("   - 消息ID匹配正确")
+        print("   - 成对删除逻辑有效")
+        print("   - 裁剪算法运行正常")
+    else:
+        print("\n⚠️ 消息ID裁剪逻辑存在问题，请检查相关代码")
+    
+    return success
+
+
+if __name__ == "__main__":
+    success = main()
+    sys.exit(0 if success else 1) 

autocoder/common/pull_requests/__init__.py

@@ -21,7 +21,8 @@ from .utils import (
     parse_git_url, detect_platform_from_repo, get_repo_info_from_path,
     get_current_branch, branch_exists, is_git_repo, get_default_remote_branch,
     ensure_branch_exists_remotely, is_main_branch, create_and_checkout_branch,
-    generate_auto_branch_name
+    generate_auto_branch_name, has_uncommitted_changes, auto_commit_changes,
+    ensure_branch_has_commits
 )
 
 
@@ -92,6 +93,13 @@ def create_pull_request(
     if source_branch == target_branch:
         raise ValidationError(f"源分支和目标分支不能相同: {source_branch}")
     
+    # 确保分支有提交，如果没有则自动提交更改
+    if not ensure_branch_has_commits(repo_path, source_branch, auto_commit=True):
+        raise ValidationError(
+            f"分支 '{source_branch}' 没有提交且无法自动提交更改。"
+            f"请确保有需要提交的更改或手动提交后再创建PR。"
+        )
+    
     # 确保源分支存在于远程仓库，如果不存在则自动推送
     if not ensure_branch_exists_remotely(repo_path, source_branch):
         raise BranchNotFoundError(

autocoder/common/pull_requests/utils.py

@@ -325,6 +325,127 @@ def validate_repo_path(repo_path: str) -> str:
     return str(path)
 
 
+def has_uncommitted_changes(repo_path: str) -> bool:
+    """检查是否有未提交的更改"""
+    try:
+        # 检查工作区是否有未暂存的更改
+        result = subprocess.run(
+            ['git', 'diff', '--quiet'],
+            cwd=repo_path,
+            capture_output=True
+        )
+        has_unstaged = result.returncode != 0
+        
+        # 检查暂存区是否有未提交的更改
+        result = subprocess.run(
+            ['git', 'diff', '--cached', '--quiet'],
+            cwd=repo_path,
+            capture_output=True
+        )
+        has_staged = result.returncode != 0
+        
+        return has_unstaged or has_staged
+    except subprocess.CalledProcessError:
+        return False
+
+
+def get_uncommitted_files(repo_path: str) -> list:
+    """获取未提交的文件列表"""
+    try:
+        result = subprocess.run(
+            ['git', 'status', '--porcelain'],
+            cwd=repo_path,
+            capture_output=True,
+            text=True,
+            check=True
+        )
+        lines = result.stdout.strip().split('\n')
+        return [line.strip() for line in lines if line.strip()]
+    except subprocess.CalledProcessError:
+        return []
+
+
+def auto_commit_changes(repo_path: str, commit_message: str = None) -> bool:
+    """自动提交所有更改"""
+    try:
+        # 检查是否有更改
+        if not has_uncommitted_changes(repo_path):
+            logger.info("没有需要提交的更改")
+            return True
+        
+        # 获取未提交的文件列表用于日志
+        uncommitted_files = get_uncommitted_files(repo_path)
+        logger.info(f"发现 {len(uncommitted_files)} 个未提交的更改")
+        
+        # 添加所有更改到暂存区
+        logger.info("正在添加所有更改到暂存区...")
+        subprocess.run(
+            ['git', 'add', '.'],
+            cwd=repo_path,
+            check=True
+        )
+        
+        # 生成提交消息
+        if not commit_message:
+            current_branch = get_current_branch(repo_path)
+            if current_branch and current_branch.startswith('ac-'):
+                commit_message = f"Auto-commit changes for PR from {current_branch}"
+            else:
+                commit_message = "Auto-commit changes for PR creation"
+        
+        # 提交更改
+        logger.info(f"正在提交更改: {commit_message}")
+        subprocess.run(
+            ['git', 'commit', '-m', commit_message],
+            cwd=repo_path,
+            check=True
+        )
+        
+        logger.info("✓ 成功提交所有更改")
+        return True
+        
+    except subprocess.CalledProcessError as e:
+        logger.error(f"自动提交失败: {e}")
+        return False
+
+
+def ensure_branch_has_commits(repo_path: str, branch_name: str, auto_commit: bool = True) -> bool:
+    """确保分支有提交，如果没有则自动提交更改"""
+    try:
+        # 检查分支是否有提交（相对于目标分支的差异）
+        result = subprocess.run(
+            ['git', 'rev-list', '--count', f'origin/main..{branch_name}'],
+            cwd=repo_path,
+            capture_output=True,
+            text=True,
+            check=True
+        )
+        commit_count = int(result.stdout.strip())
+        
+        if commit_count > 0:
+            logger.info(f"分支 '{branch_name}' 已有 {commit_count} 个提交")
+            return True
+        
+        # 如果没有提交，检查是否有未提交的更改
+        if has_uncommitted_changes(repo_path):
+            if auto_commit:
+                logger.info(f"分支 '{branch_name}' 没有提交但有未提交的更改，正在自动提交...")
+                return auto_commit_changes(repo_path)
+            else:
+                logger.warning(f"分支 '{branch_name}' 没有提交且有未提交的更改，但未启用自动提交")
+                return False
+        else:
+            logger.warning(f"分支 '{branch_name}' 没有提交也没有更改")
+            return False
+            
+    except subprocess.CalledProcessError as e:
+        logger.error(f"检查分支提交状态失败: {e}")
+        # 如果检查失败，尝试自动提交（如果有更改的话）
+        if auto_commit and has_uncommitted_changes(repo_path):
+            return auto_commit_changes(repo_path)
+        return False
+
+
 def build_pr_url(platform: PlatformType, repo_info: RepoInfo, pr_number: int) -> str:
     """构建PR的Web URL"""
     base_urls = {
@@ -341,4 +462,4 @@ def build_pr_url(platform: PlatformType, repo_info: RepoInfo, pr_number: int) ->
     if platform == PlatformType.GITLAB or platform == PlatformType.GITCODE:
         return f"{base_url}/{repo_info.full_name}/-/merge_requests/{pr_number}"
     else:
-        return f"{base_url}/{repo_info.full_name}/pull/{pr_number}"
+        return f"{base_url}/{repo_info.full_name}/pull/{pr_number}"

autocoder/common/rag_manager/rag_manager.py

@@ -1,6 +1,7 @@
 import os
 import json
-from typing import List, Dict, Any, Optional
+from typing import List, Optional
+from pathlib import Path
 from pydantic import BaseModel
 from autocoder.common import AutoCoderArgs
 from loguru import logger
@@ -16,42 +17,37 @@ class RAGConfig(BaseModel):
 
 class RAGManager:
     """RAG 管理器，用于读取和管理 RAG 服务器配置"""
-    
+
     def __init__(self, args: AutoCoderArgs):
         self.args = args
         self.configs: List[RAGConfig] = []
         self._load_configs()
-    
+
     def _load_configs(self):
         """加载 RAG 配置，优先从项目配置，然后从全局配置"""
         # 优先读取项目级别配置
-        project_config_path = os.path.join(
-            self.args.source_dir, 
-            ".auto-coder", 
-            "auto-coder.web", 
-            "rags", 
-            "rags.json"
-        )
-        
-        if os.path.exists(project_config_path):
+        base_path = Path(self.args.source_dir) if self.args.source_dir else Path.cwd()
+        project_config_path = base_path / ".auto-coder" / "auto-coder.web" / "rags" / "rags.json"
+
+        if project_config_path.exists():
             logger.info(f"正在加载项目级别 RAG 配置: {project_config_path}")
-            self._load_project_config(project_config_path)
+            self._load_project_config(str(project_config_path))
         else:
             logger.info("未找到项目级别 RAG 配置，尝试加载全局配置")
-            # 读取全局配置
-            global_config_path = os.path.expanduser("~/.auto-coder/keys/rags.json")
-            if os.path.exists(global_config_path):
+            # 读取全局配置（使用 pathlib 确保跨平台兼容性）
+            global_config_path = Path.home() / ".auto-coder" / "keys" / "rags.json"
+            if global_config_path.exists():
                 logger.info(f"正在加载全局 RAG 配置: {global_config_path}")
-                self._load_global_config(global_config_path)
+                self._load_global_config(str(global_config_path))
             else:
                 logger.warning("未找到任何 RAG 配置文件")
-    
+
     def _load_project_config(self, config_path: str):
         """加载项目级别的 RAG 配置"""
         try:
             with open(config_path, 'r', encoding='utf-8') as f:
                 config_data = json.load(f)
-            
+
             if "data" in config_data and isinstance(config_data["data"], list):
                 for item in config_data["data"]:
                     try:
@@ -67,18 +63,18 @@ class RAGManager:
                         logger.error(f"解析项目级别 RAG 配置项时出错: {e}, 配置项: {item}")
             else:
                 logger.error(f"项目级别 RAG 配置格式错误，缺少 'data' 字段或 'data' 不是列表")
-                
+
         except json.JSONDecodeError as e:
             logger.error(f"项目级别 RAG 配置文件 JSON 格式错误: {e}")
         except Exception as e:
             logger.error(f"读取项目级别 RAG 配置文件时出错: {e}")
-    
+
     def _load_global_config(self, config_path: str):
         """加载全局级别的 RAG 配置"""
         try:
             with open(config_path, 'r', encoding='utf-8') as f:
                 config_data = json.load(f)
-            
+
             if "data" in config_data and isinstance(config_data["data"], list):
                 for item in config_data["data"]:
                     try:
@@ -94,49 +90,49 @@ class RAGManager:
                         logger.error(f"解析全局 RAG 配置项时出错: {e}, 配置项: {item}")
             else:
                 logger.error(f"全局 RAG 配置格式错误，缺少 'data' 字段或 'data' 不是列表")
-                
+
         except json.JSONDecodeError as e:
             logger.error(f"全局 RAG 配置文件 JSON 格式错误: {e}")
         except Exception as e:
             logger.error(f"读取全局 RAG 配置文件时出错: {e}")
-    
+
     def get_all_configs(self) -> List[RAGConfig]:
         """获取所有 RAG 配置"""
         return self.configs
-    
+
     def get_config_by_name(self, name: str) -> Optional[RAGConfig]:
         """根据名称获取特定的 RAG 配置"""
         for config in self.configs:
             if config.name == name:
                 return config
         return None
-    
+
     def get_server_names(self) -> List[str]:
         """获取所有服务器名称列表"""
         return [config.server_name for config in self.configs]
-    
+
     def get_config_info(self) -> str:
         """获取格式化的配置信息，用于显示"""
         if not self.configs:
-            return "未找到可用的 RAG 服务器配置"
-        
+            return "### RAG_SERVER_LIST\n No available RAG server configurations found"
+
         info_lines = []
-        info_lines.append("可用的 RAG 服务器配置")
-        
+        info_lines.append("### RAG_SERVER_LIST\nAvailable RAG server configurations")
+
         for i, config in enumerate(self.configs, 1):
-            info_lines.append(f"\n{i}. 配置名称: {config.name}")
-            info_lines.append(f"   服务器地址: {config.server_name}")
-            
+            info_lines.append(f"\n{i}. Configuration name: {config.name}")
+            info_lines.append(f"   Server address: {config.server_name}")
+
             if config.description:
-                info_lines.append(f"   描述信息: {config.description}")
+                info_lines.append(f"   Description: {config.description}")
             else:
-                info_lines.append(f"   描述信息: 无")                           
-                
+                info_lines.append(f"   Description: None")
+
             if i < len(self.configs):
                 info_lines.append("-" * 30)
-                
+
         return "\n".join(info_lines)
-    
+
     def has_configs(self) -> bool:
         """检查是否有可用的配置"""
-        return len(self.configs) > 0 
+        return len(self.configs) > 0

autocoder/common/rulefiles/__init__.py

@@ -4,12 +4,64 @@
 AutoCoder 规则文件管理模块
 
 提供读取、解析和监控 AutoCoder 规则文件的功能。
+
+重构后的模块结构：
+- models/: 数据模型 (RuleFile, RuleRelevance, AlwaysApplyRuleSummary, ConditionalRulesIndex)
+- core/: 核心功能 (AutocoderRulesManager, RuleSelector)  
+- utils/: 工具函数 (parser, monitor)
+- api.py: 统一的 API 接口
 """
 
-from .autocoderrules_utils import (
+# 导入数据模型
+from .models import (
+    RuleFile,
+    RuleRelevance,
+    AlwaysApplyRuleSummary,
+    ConditionalRulesIndex,
+    InitRule,
+)
+
+# 导入核心组件
+from .core import (
+    AutocoderRulesManager,
+    RuleSelector,
+)
+
+# 导入公共 API 函数
+from .api import (
     get_rules,
+    get_parsed_rules,
+    parse_rule_file,
+    reset_rules_manager,
+    auto_select_rules,
+    get_required_and_index_rules,
+    generate_always_apply_summary,
+    generate_conditional_rules_index,
+    init_rule,
+    get_rules_for_conversation,
 )
 
 __all__ = [
+    # 数据模型
+    'RuleFile',
+    'RuleRelevance',
+    'AlwaysApplyRuleSummary',
+    'ConditionalRulesIndex',
+    'InitRule',
+    
+    # 核心组件
+    'AutocoderRulesManager',
+    'RuleSelector',
+    
+    # 公共 API 函数
     'get_rules',
+    'get_parsed_rules',
+    'parse_rule_file',
+    'reset_rules_manager',
+    'auto_select_rules',
+    'get_required_and_index_rules',
+    'generate_always_apply_summary',
+    'generate_conditional_rules_index',
+    'init_rule',
+    'get_rules_for_conversation',
 ]