jarvis-ai-assistant 0.3.30__py3-none-any.whl → 0.7.6__py3-none-any.whl

jarvis/jarvis_tools/read_code.py

@@ -1,13 +1,33 @@
 # -*- coding: utf-8 -*-
 import os
-from typing import Any, Dict
+import time
+from typing import Any, Dict, List
 
-from jarvis.jarvis_utils.output import OutputType, PrettyOutput
+from jarvis.jarvis_utils.config import get_max_input_token_count
+from jarvis.jarvis_utils.embedding import get_context_token_count
+
+# 尝试导入语言支持模块
+try:
+    from jarvis.jarvis_code_agent.code_analyzer.language_support import (
+        detect_language,
+        get_dependency_analyzer,
+    )
+    from jarvis.jarvis_code_agent.code_analyzer.structured_code import StructuredCodeExtractor
+    LANGUAGE_SUPPORT_AVAILABLE = True
+except ImportError:
+    LANGUAGE_SUPPORT_AVAILABLE = False
+    def get_dependency_analyzer(language: str):
+        return None
+    StructuredCodeExtractor = None
 
 
 class ReadCodeTool:
     name = "read_code"
-    description = "代码阅读与分析工具，用于读取源代码文件并添加行号，针对代码文件优化，提供更好的格式化输出和行号显示，适用于代码分析、审查和理解代码实现的场景"
+    description = (
+        "结构化读取源代码文件。"
+        "支持的语言按语法单元（函数、类等）读取；不支持的语言按空白行分组；"
+        "raw_mode=true 时按每20行分组读取。"
+    )
     # 工具标签
     parameters = {
         "type": "object",
@@ -20,17 +40,667 @@ class ReadCodeTool:
                         "path": {"type": "string"},
                         "start_line": {"type": "number", "default": 1},
                         "end_line": {"type": "number", "default": -1},
+                        "raw_mode": {"type": "boolean", "default": False},
                     },
                     "required": ["path"],
                 },
-                "description": "要读取的文件列表",
+                "description": "要读取的文件列表，每个文件可指定行号范围（start_line 到 end_line，-1 表示文件末尾）。raw_mode为true时按每20行分组读取（原始模式）。",
             }
         },
         "required": ["files"],
     }
-
+    
+    def _extract_syntax_units(
+        self, filepath: str, content: str, start_line: int, end_line: int
+    ) -> List[Dict[str, Any]]:
+        """提取语法单元（函数、类等）
+        
+        Args:
+            filepath: 文件路径
+            content: 文件内容
+            start_line: 起始行号
+            end_line: 结束行号
+            
+        Returns:
+            语法单元列表，每个单元包含 id, start_line, end_line, content
+        """
+        if StructuredCodeExtractor:
+            return StructuredCodeExtractor.extract_syntax_units(filepath, content, start_line, end_line)
+        return []
+    
+    def _extract_syntax_units_with_split(
+        self, filepath: str, content: str, start_line: int, end_line: int
+    ) -> List[Dict[str, Any]]:
+        """提取语法单元，然后对超过50行的单元进行二级切分：
+        1. 先按连续空白行切分大块
+        2. 如果子块仍然超过50行，再按固定行数（50行一组）切分
+        
+        Args:
+            filepath: 文件路径
+            content: 文件内容
+            start_line: 起始行号
+            end_line: 结束行号
+            
+        Returns:
+            语法单元列表，每个单元不超过50行
+        """
+        # 先获取语法单元（仅在支持语法解析的语言中才会返回非空）
+        syntax_units = self._extract_syntax_units(filepath, content, start_line, end_line)
+        
+        if not syntax_units:
+            return []
+        
+        result = []
+        for unit in syntax_units:
+            unit_line_count = unit['end_line'] - unit['start_line'] + 1
+            if unit_line_count > 50:
+                # 第一步：对大块先按空白行切分（基于 StructuredCodeExtractor）
+                blank_groups = self._extract_blank_line_groups(
+                    content, unit['start_line'], unit['end_line']
+                )
+                
+                # 如果按空白行切分失败（例如全部为空白或实现返回空），退回原始大块
+                if not blank_groups:
+                    blank_groups = [unit]
+                
+                for group in blank_groups:
+                    group_line_count = group['end_line'] - group['start_line'] + 1
+                    if group_line_count > 50:
+                        # 第二步：对子块中仍然超过50行的部分，按每50行固定切分
+                        sub_groups = self._extract_line_groups(
+                            content, group['start_line'], group['end_line'], group_size=50
+                        )
+                        result.extend(sub_groups)
+                    else:
+                        # 经过空白行切分得到的中等大小块，直接加入结果
+                        result.append(group)
+            else:
+                # 如果单元不超过50行，直接添加
+                result.append(unit)
+        
+        return result
+    
+    def _extract_blank_line_groups(
+        self, content: str, start_line: int, end_line: int
+    ) -> List[Dict[str, Any]]:
+        """按空白行分组提取内容（委托给StructuredCodeExtractor）"""
+        if StructuredCodeExtractor:
+            return StructuredCodeExtractor.extract_blank_line_groups(content, start_line, end_line)
+        return []
+    
+    def _extract_blank_line_groups_with_split(
+        self, content: str, start_line: int, end_line: int
+    ) -> List[Dict[str, Any]]:
+        """先按空白行分组，然后对超过20行的块再按每20行分割
+        
+        Args:
+            content: 文件内容
+            start_line: 起始行号
+            end_line: 结束行号
+            
+        Returns:
+            分组列表，每个分组包含 id, start_line, end_line, content
+        """
+        # 先获取空白行分组
+        blank_line_groups = self._extract_blank_line_groups(content, start_line, end_line)
+        
+        if not blank_line_groups:
+            return []
+        
+        result = []
+        for group in blank_line_groups:
+            group_line_count = group['end_line'] - group['start_line'] + 1
+            if group_line_count > 20:
+                # 如果块超过20行，按每20行分割
+                sub_groups = self._extract_line_groups(
+                    content, group['start_line'], group['end_line'], group_size=20
+                )
+                result.extend(sub_groups)
+            else:
+                # 如果块不超过20行，直接添加
+                result.append(group)
+        
+        return result
+    
+    def _extract_line_groups(
+        self, content: str, start_line: int, end_line: int, group_size: int = 20
+    ) -> List[Dict[str, Any]]:
+        """按行号分组提取内容（委托给StructuredCodeExtractor）"""
+        if StructuredCodeExtractor:
+            return StructuredCodeExtractor.extract_line_groups(content, start_line, end_line, group_size)
+        return []
+    
+    def _ensure_unique_ids(self, units: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
+        """确保单元列表中所有id唯一（委托给StructuredCodeExtractor）"""
+        if StructuredCodeExtractor:
+            return StructuredCodeExtractor.ensure_unique_ids(units)
+        return units
+    
+    def _extract_imports(self, filepath: str, content: str, start_line: int, end_line: int) -> List[Dict[str, Any]]:
+        """提取文件的导入/包含语句作为结构化单元（委托给StructuredCodeExtractor）"""
+        if StructuredCodeExtractor:
+            return StructuredCodeExtractor.extract_imports(filepath, content, start_line, end_line)
+        return []
+    
+    def _create_import_unit(self, import_group: List[Dict[str, Any]]) -> Dict[str, Any]:
+        """创建导入语句单元（委托给StructuredCodeExtractor）"""
+        if StructuredCodeExtractor:
+            return StructuredCodeExtractor.create_import_unit(import_group)
+        return {}
+    
+    def _format_structured_output(
+        self, filepath: str, units: List[Dict[str, Any]], total_lines: int, agent: Any = None
+    ) -> str:
+        """格式化结构化输出
+        
+        Args:
+            filepath: 文件路径
+            units: 语法单元或行号分组列表（已包含导入语句单元）
+            total_lines: 文件总行数
+            agent: Agent实例，用于从缓存中获取block_id
+            
+        Returns:
+            格式化后的输出字符串
+        """
+        # 文件开始分界符
+        output_lines = [
+            "=" * 80,
+            f"🔍 文件: {filepath}",
+            f"📄 总行数: {total_lines}",
+            f"📦 结构化单元数: {len(units)}",
+            "=" * 80,
+            "",
+        ]
+        
+        # 为每个单元分配block-id
+        # 如果unit已经有block_id（从缓存中获取），直接使用；否则按顺序生成
+        for idx, unit in enumerate(units, start=1):
+            # 如果unit已经有block_id，直接使用（在生成structured_units时已分配）
+            block_id = unit.get('block_id')
+            if not block_id:
+                # 否则按顺序生成临时id
+                block_id = f"block-{idx}"
+            # 显示id
+            output_lines.append(f"[id:{block_id}]")
+            # 添加内容，保持原有缩进，并为每行添加行号
+            content = unit.get('content', '')
+            if content:
+                # 获取单元的起始行号
+                start_line = unit.get('start_line', 1)
+                # 将内容按行分割
+                content_lines = content.split('\n')
+                # 为每一行添加行号（右对齐，4位，不足补空格）
+                numbered_lines = []
+                current_line = start_line
+                for line in content_lines:
+                    # 行号右对齐，占4位
+                    line_number_str = f"{current_line:4d}"
+                    numbered_lines.append(f"{line_number_str}:{line}")
+                    current_line += 1
+                # 将带行号的内容添加到输出
+                output_lines.append('\n'.join(numbered_lines))
+            # 块结束分界符
+            output_lines.append("-" * 80)
+            output_lines.append("")  # 单元之间空行分隔
+        
+        # 文件结束分界符
+        output_lines.append("=" * 80)
+        output_lines.append("")
+        
+        return '\n'.join(output_lines)
+    
+    def _get_file_cache(self, agent: Any, filepath: str) -> Dict[str, Any]:
+        """获取文件的缓存信息
+        
+        Args:
+            agent: Agent实例
+            filepath: 文件路径
+            
+        Returns:
+            缓存信息字典，如果不存在则返回None
+        """
+        if not agent:
+            return None
+        
+        cache = agent.get_user_data("read_code_cache")
+        if not cache:
+            return None
+        
+        abs_path = os.path.abspath(filepath)
+        return cache.get(abs_path)
+    
+    def _get_blocks_from_cache(self, cache_info: Dict[str, Any], start_line: int, end_line: int) -> List[Dict[str, Any]]:
+        """从缓存中获取对应范围的blocks
+        
+        Args:
+            cache_info: 缓存信息
+            start_line: 起始行号（1-based）
+            end_line: 结束行号（1-based，-1表示文件末尾）
+            
+        Returns:
+            blocks列表，每个block包含block_id和content
+        """
+        if not cache_info or "id_list" not in cache_info or "blocks" not in cache_info:
+            return []
+        
+        id_list = cache_info.get("id_list", [])
+        blocks = cache_info.get("blocks", {})
+        result = []
+        
+        # 如果end_line是-1，表示文件末尾，需要先计算文件总行数
+        if end_line == -1:
+            # 先遍历所有blocks计算总行数
+            # 注意：块内容不包含末尾换行符，块之间需要添加换行符
+            total_lines = 0
+            for idx, block_id in enumerate(id_list):
+                block_data = blocks.get(block_id)
+                if block_data:
+                    block_content = block_data.get("content", "")
+                    if block_content:
+                        # 块内容中的换行符数量 + 1 = 行数
+                        block_line_count = block_content.count('\n') + 1
+                        total_lines += block_line_count
+                        # 如果不是最后一个块，块之间有一个换行符分隔（已计入下一个块的第一行）
+                        # 所以不需要额外添加
+            end_line = total_lines
+        
+        # 通过前面blocks的内容推算每个block的行号范围
+        # 注意：块内容不包含末尾换行符，块之间需要添加换行符
+        current_line = 1  # 从第1行开始
+        
+        for idx, block_id in enumerate(id_list):
+            block_data = blocks.get(block_id)
+            if not block_data:
+                continue
+            block_content = block_data.get("content", "")
+            if not block_content:
+                continue
+            
+            # 计算这个block的行数
+            # 块内容中的换行符数量 + 1 = 行数（因为块内容不包含末尾换行符）
+            block_line_count = block_content.count('\n') + 1
+            
+            block_start_line = current_line
+            block_end_line = current_line + block_line_count - 1
+            
+            # block与请求范围有重叠就包含
+            if block_end_line >= start_line and block_start_line <= end_line:
+                result.append({
+                    "block_id": block_id,
+                    "content": block_content,
+                    "start_line": block_start_line,
+                })
+            
+            # 更新当前行号
+            # 块之间有一个换行符分隔，所以下一个块从 block_end_line + 1 开始
+            current_line = block_end_line + 1
+            
+            # 如果已经超过请求的结束行，可以提前退出
+            if block_start_line > end_line:
+                break
+        
+        return result
+    
+    def _convert_units_to_sequential_ids(self, units: List[Dict[str, Any]], full_content: str = None) -> Dict[str, Any]:
+        """将单元列表转换为缓存格式（id_list和blocks字典）
+        
+        按照行号范围分割文件，不区分语法单元，确保完美恢复。
+        
+        Args:
+            units: 结构化单元列表，每个单元包含 id, start_line, end_line, content
+            full_content: 完整的文件内容（可选），用于确保块之间的空白行也被包含
+            
+        Returns:
+            包含 id_list 和 blocks 的字典：
+            - id_list: 有序的id列表，如 ["block-1", "block-2", "block-3"]
+            - blocks: id到块信息的字典，如 {"block-1": {"content": "..."}, ...}
+        """
+        if not full_content or not units:
+            # 没有完整内容，直接使用原始的content
+            sorted_original = sorted(units, key=lambda u: u.get('start_line', 0))
+            id_list = []
+            blocks = {}
+            for unit in sorted_original:
+                block_id = f"block-{len(id_list) + 1}"  # block-1, block-2, ...
+                id_list.append(block_id)
+                content = unit.get('content', '')
+                # 去掉块末尾的换行符
+                if content.endswith('\n'):
+                    content = content[:-1]
+                blocks[block_id] = {
+                    "content": content,
+                }
+            return {
+                "id_list": id_list,
+                "blocks": blocks,
+                "file_ends_with_newline": False,  # 无法确定，默认False
+            }
+        
+        # 收集所有单元的开始行号作为分割点
+        # 关键：直接使用每个单元的start_line，不合并范围，保留语法单元边界
+        split_points_set = {1}  # 从第1行开始
+        for unit in units:
+            start_line = unit.get('start_line', 1)
+            if start_line > 0:
+                split_points_set.add(start_line)
+        
+        if not split_points_set:
+            # 没有有效的分割点，返回空列表
+            return {"id_list": [], "blocks": {}, "file_ends_with_newline": False}
+        
+        # 按照每个单元的开始行作为分割点，连续分割文件内容
+        # 每个块包含从当前分割点到下一个分割点之前的所有内容
+        # 关键：直接按行号范围从原始内容中提取，确保完美恢复（包括文件末尾的换行符和所有空白行）
+        # 使用 split('\n') 分割，然后手动为每行添加换行符（除了最后一行，根据原始文件决定）
+        lines = full_content.split('\n')
+        result_units = []
+        
+        # 排序分割点
+        split_points = sorted(split_points_set)
+        split_points.append(len(lines) + 1)  # 文件末尾
+        
+        # 按照分割点连续分割文件
+        # 注意：如果文件以换行符结尾，split('\n')会在末尾产生一个空字符串
+        # 我们需要正确处理这种情况
+        file_ends_with_newline = full_content.endswith('\n')
+        
+        for idx in range(len(split_points) - 1):
+            start_line = split_points[idx]  # 1-based
+            next_start_line = split_points[idx + 1]  # 1-based
+            
+            # 提取从当前分割点到下一个分割点之前的所有内容
+            unit_start_idx = max(0, start_line - 1)  # 0-based索引
+            unit_end_idx = min(len(lines) - 1, next_start_line - 2)  # 0-based索引，下一个分割点之前
+            
+            # 确保索引有效
+            if unit_start_idx <= unit_end_idx:
+                # 提取行并重新组合，确保保留所有换行符
+                extracted_lines = lines[unit_start_idx:unit_end_idx + 1]
+                
+                # 重新组合：每行后面添加换行符
+                # 对于非最后一个块，最后一行也需要换行符，因为下一个块从下一行开始
+                # 对于最后一个块，根据原始文件是否以换行符结尾来决定
+                full_unit_content_parts = []
+                is_last_block = (idx == len(split_points) - 2)
+                
+                for i, line in enumerate(extracted_lines):
+                    if i < len(extracted_lines) - 1:
+                        # 不是最后一行，添加换行符
+                        full_unit_content_parts.append(line + '\n')
+                    else:
+                        # 最后一行
+                        if not is_last_block:
+                            # 非最后一个块：最后一行必须添加换行符，因为下一个块从下一行开始
+                            # 这样可以保留块之间的空白行
+                            full_unit_content_parts.append(line + '\n')
+                        else:
+                            # 最后一个块：需要特殊处理
+                            # 如果文件以换行符结尾，且最后一行是空字符串（来自split('\n')的副作用），
+                            # 且不是唯一的一行，那么前面的行已经输出了换行符，这里不需要再输出
+                            if file_ends_with_newline and line == '' and len(extracted_lines) > 1:
+                                # 最后一行是空字符串且来自trailing newline，且不是唯一的一行
+                                # 前面的行已经输出了换行符，所以这里不需要再输出任何内容
+                                # 空字符串表示不输出任何内容
+                                full_unit_content_parts.append('')
+                            elif file_ends_with_newline:
+                                # 文件以换行符结尾，最后一行需要换行符
+                                full_unit_content_parts.append(line + '\n')
+                            else:
+                                # 文件不以换行符结尾
+                                full_unit_content_parts.append(line)
+                
+                full_unit_content = ''.join(full_unit_content_parts)
+                
+                # 去掉块末尾的换行符（存储时去掉，恢复时再添加）
+                if full_unit_content.endswith('\n'):
+                    full_unit_content = full_unit_content[:-1]
+                
+                block_id = f"block-{len(result_units) + 1}"  # block-1, block-2, ...
+                result_units.append({
+                    "id": block_id,
+                    "content": full_unit_content,
+                })
+        
+        # 转换为 id_list 和 blocks 格式
+        id_list = [unit["id"] for unit in result_units]
+        blocks = {
+            unit["id"]: {
+                "content": unit["content"],
+            }
+            for unit in result_units
+        }
+        
+        # 保存文件是否以换行符结尾的信息（用于恢复时正确处理）
+        file_ends_with_newline = full_content.endswith('\n')
+        
+        return {
+            "id_list": id_list,
+            "blocks": blocks,
+            "file_ends_with_newline": file_ends_with_newline,
+        }
+    
+    def _save_file_cache(
+        self, agent: Any, filepath: str, units: List[Dict[str, Any]], 
+        total_lines: int, file_mtime: float, full_content: str = None
+    ) -> None:
+        """保存文件的结构化信息到缓存
+        
+        Args:
+            agent: Agent实例
+            filepath: 文件路径
+            units: 结构化单元列表
+            total_lines: 文件总行数
+            file_mtime: 文件修改时间
+            full_content: 完整的文件内容（可选），用于确保块之间的空白行也被包含
+        """
+        if not agent:
+            return
+        
+        cache = agent.get_user_data("read_code_cache")
+        if not cache:
+            cache = {}
+            agent.set_user_data("read_code_cache", cache)
+        
+        abs_path = os.path.abspath(filepath)
+        
+        # 转换为 id_list 和 blocks 格式
+        cache_data = self._convert_units_to_sequential_ids(units, full_content)
+        
+        cache[abs_path] = {
+            "id_list": cache_data["id_list"],
+            "blocks": cache_data["blocks"],
+            "total_lines": total_lines,
+            "read_time": time.time(),
+            "file_mtime": file_mtime,
+            "file_ends_with_newline": cache_data.get("file_ends_with_newline", False),
+        }
+        agent.set_user_data("read_code_cache", cache)
+    
+    def _is_cache_valid(self, cache_info: Dict[str, Any], filepath: str) -> bool:
+        """检查缓存是否有效
+        
+        Args:
+            cache_info: 缓存信息字典
+            filepath: 文件路径
+            
+        Returns:
+            True表示缓存有效，False表示缓存无效
+        """
+        if not cache_info:
+            return False
+        
+        try:
+            # 检查文件是否存在
+            if not os.path.exists(filepath):
+                return False
+            
+            # 检查文件修改时间是否变化
+            current_mtime = os.path.getmtime(filepath)
+            cached_mtime = cache_info.get("file_mtime")
+            
+            if cached_mtime is None or abs(current_mtime - cached_mtime) > 0.1:  # 允许0.1秒的误差
+                return False
+            
+            # 检查缓存数据结构是否完整
+            if "id_list" not in cache_info or "blocks" not in cache_info or "total_lines" not in cache_info:
+                return False
+            
+            return True
+        except Exception:
+            return False
+    
+    def _restore_file_from_cache(self, cache_info: Dict[str, Any]) -> str:
+        """从缓存恢复文件内容
+        
+        Args:
+            cache_info: 缓存信息字典
+            
+        Returns:
+            恢复的文件内容字符串（与原始文件内容完全一致）
+        """
+        if not cache_info:
+            return ""
+        
+        # 按照 id_list 的顺序恢复
+        id_list = cache_info.get("id_list", [])
+        blocks = cache_info.get("blocks", {})
+        file_ends_with_newline = cache_info.get("file_ends_with_newline", False)
+        
+        result = []
+        for idx, block_id in enumerate(id_list):
+            block = blocks.get(block_id)
+            if block:
+                content = block.get('content', '')
+                if content:
+                    result.append(content)
+                    # 在块之间添加换行符（最后一个块后面根据文件是否以换行符结尾决定）
+                    is_last_block = (idx == len(id_list) - 1)
+                    if is_last_block:
+                        # 最后一个块：如果文件以换行符结尾，添加换行符
+                        if file_ends_with_newline:
+                            result.append('\n')
+                    else:
+                        # 非最后一个块：在块之间添加换行符
+                        result.append('\n')
+        
+        return ''.join(result) if result else ""
+    
+    def _estimate_structured_tokens(
+        self, filepath: str, content: str, start_line: int, end_line: int, total_lines: int, raw_mode: bool = False
+    ) -> int:
+        """估算结构化输出的token数
+        
+        Args:
+            filepath: 文件路径
+            content: 文件内容
+            start_line: 起始行号
+            end_line: 结束行号
+            total_lines: 文件总行数
+            
+        Returns:
+            估算的token数
+        """
+        try:
+            if raw_mode:
+                # 原始模式：按每20行分组计算token
+                line_groups = self._extract_line_groups(content, start_line, end_line, group_size=20)
+                if line_groups:
+                    import_units = self._extract_imports(filepath, content, start_line, end_line)
+                    all_units = import_units + line_groups[:1]
+                    # 确保id唯一
+                    all_units = self._ensure_unique_ids(all_units)
+                    # 按行号排序
+                    all_units.sort(key=lambda u: u['start_line'])
+                    sample_output = self._format_structured_output(filepath, all_units, total_lines)
+                    if len(line_groups) > 1:
+                        group_tokens = get_context_token_count(sample_output)
+                        return group_tokens * len(line_groups)
+                    else:
+                        return get_context_token_count(sample_output)
+            else:
+                # 尝试提取语法单元（确保每个单元不超过50行）
+                syntax_units = self._extract_syntax_units_with_split(filepath, content, start_line, end_line)
+                
+                if syntax_units:
+                    # 使用语法单元结构化输出格式计算token
+                    import_units = self._extract_imports(filepath, content, start_line, end_line)
+                    all_units = import_units + syntax_units[:1]
+                    # 确保id唯一
+                    all_units = self._ensure_unique_ids(all_units)
+                    # 按行号排序
+                    all_units.sort(key=lambda u: u['start_line'])
+                    sample_output = self._format_structured_output(filepath, all_units, total_lines)
+                    if len(syntax_units) > 1:
+                        unit_tokens = get_context_token_count(sample_output)
+                        return unit_tokens * len(syntax_units)
+                    else:
+                        return get_context_token_count(sample_output)
+                else:
+                    # 使用空白行分组格式计算token（不支持语言时）
+                    # 先按空行分割，然后对超过20行的块再按每20行分割
+                    line_groups = self._extract_blank_line_groups_with_split(content, start_line, end_line)
+                    if line_groups:
+                        import_units = self._extract_imports(filepath, content, start_line, end_line)
+                        all_units = import_units + line_groups[:1]
+                        # 确保id唯一
+                        all_units = self._ensure_unique_ids(all_units)
+                        # 按行号排序
+                        all_units.sort(key=lambda u: u['start_line'])
+                        sample_output = self._format_structured_output(filepath, all_units, total_lines)
+                        if len(line_groups) > 1:
+                            group_tokens = get_context_token_count(sample_output)
+                            return group_tokens * len(line_groups)
+                        else:
+                            return get_context_token_count(sample_output)
+                    else:
+                        # 回退到原始格式计算
+                        lines = content.split('\n')
+                        selected_lines = lines[start_line - 1:end_line]
+                        numbered_content = "".join(f"{i:5d}:{line}\n" for i, line in enumerate(selected_lines, start=start_line))
+                        return get_context_token_count(numbered_content)
+        except Exception:
+            # 如果估算失败，使用简单的行号格式估算
+            lines = content.split('\n')
+            selected_lines = lines[start_line - 1:end_line]
+            numbered_content = "".join(f"{i:5d}:{line}\n" for i, line in enumerate(selected_lines, start=start_line))
+            return get_context_token_count(numbered_content)
+    
+    def _get_max_token_limit(self, agent: Any = None) -> int:
+        """获取基于剩余token数量的token限制
+        
+        Args:
+            agent: Agent实例，用于获取模型和剩余token数量
+            
+        Returns:
+            int: 允许的最大token数（剩余token的2/3，或至少保留1/3剩余token）
+        """
+        try:
+            # 优先使用剩余token数量
+            if agent and hasattr(agent, "model"):
+                try:
+                    remaining_tokens = agent.model.get_remaining_token_count()
+                    # 使用剩余token的2/3作为限制，保留1/3作为安全余量
+                    limit_tokens = int(remaining_tokens * 2 / 3)
+                    # 确保至少返回一个合理的值
+                    if limit_tokens > 0:
+                        return limit_tokens
+                except Exception:
+                    pass
+            
+            # 回退方案：使用输入窗口的2/3
+            model_group = None
+            if agent:
+                model_group = getattr(agent, "model_group", None)
+            
+            max_input_tokens = get_max_input_token_count(model_group)
+            # 计算2/3限制的token数
+            limit_tokens = int(max_input_tokens * 2 / 3)
+            return limit_tokens
+        except Exception:
+            # 如果获取失败，使用默认值（假设32000 token，2/3是21333）
+            return 21333
     def _handle_single_file(
-        self, filepath: str, start_line: int = 1, end_line: int = -1, agent: Any = None
+        self, filepath: str, start_line: int = 1, end_line: int = -1, agent: Any = None, raw_mode: bool = False
     ) -> Dict[str, Any]:
         """处理单个文件的读取操作
 
@@ -38,6 +708,7 @@ class ReadCodeTool:
             filepath (str): 文件路径
             start_line (int): 起始行号，默认为1
             end_line (int): 结束行号，默认为-1表示文件末尾
+            agent: Agent实例，用于获取上下文管理器
 
         Returns:
             Dict[str, Any]: 包含成功状态、输出内容和错误信息的字典
@@ -62,10 +733,9 @@ class ReadCodeTool:
                 }
 
             # 读取文件内容
+            # 第一遍流式读取，仅统计总行数，避免一次性读入内存
             with open(abs_path, "r", encoding="utf-8", errors="ignore") as f:
-                lines = f.readlines()
-
-            total_lines = len(lines)
+                total_lines = sum(1 for _ in f)
 
             # 处理空文件情况
             if total_lines == 0:
@@ -99,21 +769,220 @@ class ReadCodeTool:
                     "stderr": f"无效的行范围 [{start_line}-{end_line}] (总行数: {total_lines})",
                 }
 
-            # 添加行号并构建输出内容
-            selected_lines = lines[start_line - 1 : end_line]
-            numbered_content = "".join(
-                [
-                    f"{i:4d}:{line}"
-                    for i, line in enumerate(selected_lines, start=start_line)
-                ]
-            )
+            # 获取文件修改时间
+            file_mtime = os.path.getmtime(abs_path)
+            
+            # 检查缓存是否有效
+            cache_info = self._get_file_cache(agent, abs_path)
+            use_cache = self._is_cache_valid(cache_info, abs_path)
 
-            # 构建输出格式
-            output = (
-                f"\n🔍 文件: {abs_path}\n"
-                f"📄 原始行号: {start_line}-{end_line} (共{total_lines}行) \n\n"
-                f"{numbered_content}\n\n"
-            )
+            # 读取完整文件内容用于语法分析和token计算
+            if use_cache:
+                # 从缓存恢复文件内容
+                full_content = self._restore_file_from_cache(cache_info)
+                # 如果恢复失败，重新读取文件
+                if not full_content:
+                    with open(abs_path, "r", encoding="utf-8", errors="ignore") as f:
+                        full_content = f.read()
+            else:
+                # 读取文件内容
+                with open(abs_path, "r", encoding="utf-8", errors="ignore") as f:
+                    full_content = f.read()
+            
+            # 读取要读取的行范围内容
+            selected_content_lines = []
+            lines = full_content.split('\n')
+            for i in range(start_line - 1, min(end_line, len(lines))):
+                selected_content_lines.append(lines[i])
+            
+            # 估算结构化输出的token数
+            content_tokens = self._estimate_structured_tokens(abs_path, full_content, start_line, end_line, total_lines, raw_mode)
+            
+            max_token_limit = self._get_max_token_limit(agent)
+            
+            # 检查单文件读取token数是否超过2/3限制
+            if content_tokens > max_token_limit:
+                read_lines = end_line - start_line + 1
+                
+                # 计算安全读取的行数 (按比例缩减)
+                safe_lines = int((max_token_limit / content_tokens) * read_lines)
+                safe_lines = max(1, min(safe_lines, read_lines))
+                safe_end_line = start_line + safe_lines - 1
+                
+                # 读取安全范围内的内容
+                selected_content_lines = []
+                for i in range(start_line - 1, min(safe_end_line, len(lines))):
+                    selected_content_lines.append(lines[i])
+                
+                # 构造部分读取结果
+                partial_content = '\n'.join(selected_content_lines)
+                
+                return {
+                    "success": True,
+                    "stdout": (
+                        f"⚠️ 警告: 仅读取前{safe_lines}行 (共{read_lines}行)，因为内容超出限制\n"
+                        f"📊 实际读取范围: {start_line}-{safe_end_line} (原请求范围: {start_line}-{end_line})\n\n"
+                        f"{partial_content}\n\n"
+                        f"💡 建议:\n"
+                        f"   1. 如需继续读取，请使用:\n"
+                        f"      start_line={safe_end_line + 1}&end_line={end_line}\n"
+                        f"   2. 需要读取全部内容? 请缩小行范围或分批读取"
+                    ),
+                    "stderr": (
+                        f"原始请求范围 {start_line}-{end_line} 超过token限制 "
+                        f"({content_tokens}/{max_token_limit} tokens)"
+                    ),
+                }
+
+            # 生成整个文件的结构化信息（用于缓存）
+            # 提取整个文件的导入/包含语句
+            full_import_units = self._extract_imports(abs_path, full_content, 1, total_lines)
+            
+            # 生成整个文件的结构化单元
+            full_structured_units = None
+            
+            if raw_mode:
+                # 原始读取模式：按每20行分组（整个文件）
+                full_line_groups = self._extract_line_groups(full_content, 1, total_lines, group_size=20)
+                # 合并导入单元和行号分组
+                full_all_units = full_import_units + full_line_groups
+                # 确保id唯一
+                full_all_units = self._ensure_unique_ids(full_all_units)
+                # 按行号排序
+                full_all_units.sort(key=lambda u: u['start_line'])
+                full_structured_units = full_all_units
+            else:
+                # 尝试提取整个文件的语法单元（确保每个单元不超过50行）
+                full_syntax_units = self._extract_syntax_units_with_split(abs_path, full_content, 1, total_lines)
+                
+                # 检测语言类型
+                if LANGUAGE_SUPPORT_AVAILABLE:
+                    try:
+                        detect_language(abs_path)
+                    except Exception:
+                        pass
+                
+                if full_syntax_units:
+                    # 合并导入单元和语法单元
+                    full_all_units = full_import_units + full_syntax_units
+                    # 确保id唯一
+                    full_all_units = self._ensure_unique_ids(full_all_units)
+                    # 按行号排序
+                    full_all_units.sort(key=lambda u: u['start_line'])
+                    full_structured_units = full_all_units
+                else:
+                    # 使用空白行分组结构化输出（不支持语言时）
+                    # 先按空行分割，然后对超过20行的块再按每20行分割（整个文件）
+                    full_line_groups = self._extract_blank_line_groups_with_split(full_content, 1, total_lines)
+                    # 合并导入单元和行号分组
+                    full_all_units = full_import_units + full_line_groups
+                    # 确保id唯一
+                    full_all_units = self._ensure_unique_ids(full_all_units)
+                    # 按行号排序
+                    full_all_units.sort(key=lambda u: u['start_line'])
+                    full_structured_units = full_all_units
+            
+            # 保存整个文件的结构化信息到缓存
+            if full_structured_units is not None:
+                self._save_file_cache(agent, abs_path, full_structured_units, total_lines, file_mtime, full_content)
+            
+            # 如果缓存有效，直接使用缓存中的blocks输出
+            if agent:
+                cache_info = self._get_file_cache(agent, abs_path)
+                if cache_info and self._is_cache_valid(cache_info, abs_path):
+                    # 直接从缓存中获取对应范围的blocks
+                    cached_blocks = self._get_blocks_from_cache(cache_info, start_line, end_line)
+                    if cached_blocks:
+                        # 转换为units格式（用于输出），保留真实的文件起始行号
+                        structured_units = []
+                        for block in cached_blocks:
+                            structured_units.append({
+                                "block_id": block["block_id"],
+                                "content": block["content"],
+                                "start_line": block.get("start_line", 1),
+                            })
+                        output = self._format_structured_output(abs_path, structured_units, total_lines, agent)
+                    else:
+                        output = ""
+                else:
+                    # 缓存无效，重新提取units
+                    # 提取请求范围的结构化单元（用于输出）
+                    import_units = self._extract_imports(abs_path, full_content, start_line, end_line)
+                    
+                    # 确定使用的结构化单元（语法单元或行号分组）
+                    structured_units = None
+                    
+                    if raw_mode:
+                        # 原始读取模式：按每20行分组
+                        line_groups = self._extract_line_groups(full_content, start_line, end_line, group_size=20)
+                        # 合并导入单元和行号分组
+                        all_units = import_units + line_groups
+                        # 确保id唯一
+                        all_units = self._ensure_unique_ids(all_units)
+                        # 按行号排序，所有单元按在文件中的实际位置排序
+                        all_units.sort(key=lambda u: u['start_line'])
+                        structured_units = all_units
+                    else:
+                        # 尝试提取语法单元（结构化读取，full_content 已在上面读取，确保每个单元不超过50行）
+                        syntax_units = self._extract_syntax_units_with_split(abs_path, full_content, start_line, end_line)
+                        
+                        if syntax_units:
+                            # 合并导入单元和语法单元
+                            all_units = import_units + syntax_units
+                            # 确保id唯一
+                            all_units = self._ensure_unique_ids(all_units)
+                            # 按行号排序，所有单元按在文件中的实际位置排序
+                            all_units.sort(key=lambda u: u['start_line'])
+                            structured_units = all_units
+                        else:
+                            # 使用空白行分组结构化输出（不支持语言时）
+                            # 先按空行分割，然后对超过20行的块再按每20行分割
+                            line_groups = self._extract_blank_line_groups_with_split(full_content, start_line, end_line)
+                            # 合并导入单元和行号分组
+                            all_units = import_units + line_groups
+                            # 确保id唯一
+                            all_units = self._ensure_unique_ids(all_units)
+                            # 按行号排序，所有单元按在文件中的实际位置排序
+                            all_units.sort(key=lambda u: u['start_line'])
+                            structured_units = all_units
+                    
+                    if structured_units:
+                        output = self._format_structured_output(abs_path, structured_units, total_lines, agent)
+                    else:
+                        output = ""
+            else:
+                # 没有agent，无法使用缓存，重新提取units
+                import_units = self._extract_imports(abs_path, full_content, start_line, end_line)
+                
+                if raw_mode:
+                    line_groups = self._extract_line_groups(full_content, start_line, end_line, group_size=20)
+                    all_units = import_units + line_groups
+                    all_units = self._ensure_unique_ids(all_units)
+                    all_units.sort(key=lambda u: u['start_line'])
+                    structured_units = all_units
+                else:
+                    syntax_units = self._extract_syntax_units_with_split(abs_path, full_content, start_line, end_line)
+                    if syntax_units:
+                        all_units = import_units + syntax_units
+                        all_units = self._ensure_unique_ids(all_units)
+                        all_units.sort(key=lambda u: u['start_line'])
+                        structured_units = all_units
+                    else:
+                        line_groups = self._extract_blank_line_groups_with_split(full_content, start_line, end_line)
+                        all_units = import_units + line_groups
+                        all_units = self._ensure_unique_ids(all_units)
+                        all_units.sort(key=lambda u: u['start_line'])
+                        structured_units = all_units
+                
+                if structured_units:
+                    output = self._format_structured_output(abs_path, structured_units, total_lines, agent)
+                else:
+                    output = ""
+
+            # 尝试获取并附加上下文信息
+            context_info = self._get_file_context(abs_path, start_line, end_line, agent)
+            if context_info:
+                output += context_info
 
             if agent:
                 files = agent.get_user_data("files")
@@ -126,9 +995,270 @@ class ReadCodeTool:
             return {"success": True, "stdout": output, "stderr": ""}
 
         except Exception as e:
-            PrettyOutput.print(str(e), OutputType.ERROR)
+            print(f"❌ {str(e)}")
             return {"success": False, "stdout": "", "stderr": f"文件读取失败: {str(e)}"}
 
+    def _handle_merged_ranges(
+        self, filepath: str, requests: List[Dict], agent: Any = None
+    ) -> Dict[str, Any]:
+        """处理同一文件的多个范围请求，合并后去重
+        
+        Args:
+            filepath: 文件绝对路径
+            requests: 范围请求列表，每个请求包含 start_line, end_line, raw_mode
+            agent: Agent实例
+            
+        Returns:
+            Dict[str, Any]: 包含成功状态、输出内容和错误信息的字典
+        """
+        try:
+            # 文件存在性检查
+            if not os.path.exists(filepath):
+                return {
+                    "success": False,
+                    "stdout": "",
+                    "stderr": f"文件不存在: {filepath}",
+                }
+            
+            # 读取文件内容
+            with open(filepath, "r", encoding="utf-8", errors="ignore") as f:
+                full_content = f.read()
+            
+            total_lines = len(full_content.split('\n'))
+            if total_lines == 0:
+                return {
+                    "success": True,
+                    "stdout": f"\n🔍 文件: {filepath}\n📄 文件为空 (0行)\n",
+                    "stderr": "",
+                }
+            
+            # 先确保缓存存在（通过读取整个文件建立缓存）
+            first_request = requests[0]
+            self._handle_single_file(
+                filepath, 1, -1, agent, first_request.get("raw_mode", False)
+            )
+            
+            # 获取缓存
+            cache_info = self._get_file_cache(agent, filepath)
+            if not cache_info or not self._is_cache_valid(cache_info, filepath):
+                # 缓存无效，使用合并范围的方式去重
+                # 合并所有范围，计算最小起始行和最大结束行
+                min_start = float('inf')
+                max_end = 0
+                raw_mode = False
+                for req in requests:
+                    start_line = req.get("start_line", 1)
+                    end_line = req.get("end_line", -1)
+                    raw_mode = raw_mode or req.get("raw_mode", False)
+                    
+                    # 处理特殊值
+                    if end_line == -1:
+                        end_line = total_lines
+                    else:
+                        end_line = max(1, min(end_line, total_lines)) if end_line >= 0 else total_lines + end_line + 1
+                    start_line = max(1, min(start_line, total_lines)) if start_line >= 0 else total_lines + start_line + 1
+                    
+                    min_start = min(min_start, start_line)
+                    max_end = max(max_end, end_line)
+                
+                # 用合并后的范围读取一次，自然就去重了
+                result = self._handle_single_file(
+                    filepath, int(min_start), int(max_end), agent, raw_mode
+                )
+                return result
+            
+            # 收集所有范围覆盖的块ID（去重）
+            seen_block_ids = set()
+            merged_blocks = []
+            
+            for req in requests:
+                start_line = req.get("start_line", 1)
+                end_line = req.get("end_line", -1)
+                
+                # 处理特殊值
+                if end_line == -1:
+                    end_line = total_lines
+                else:
+                    end_line = max(1, min(end_line, total_lines)) if end_line >= 0 else total_lines + end_line + 1
+                start_line = max(1, min(start_line, total_lines)) if start_line >= 0 else total_lines + start_line + 1
+                
+                # 从缓存获取对应范围的块
+                cached_blocks = self._get_blocks_from_cache(cache_info, start_line, end_line)
+                for block in cached_blocks:
+                    block_id = block["block_id"]
+                    if block_id not in seen_block_ids:
+                        seen_block_ids.add(block_id)
+                        merged_blocks.append(block)
+            
+            # 按block_id排序（block-1, block-2, ...）
+            def extract_block_num(block):
+                block_id = block.get("block_id", "block-0")
+                try:
+                    return int(block_id.split("-")[1])
+                except (IndexError, ValueError):
+                    return 0
+            
+            merged_blocks.sort(key=extract_block_num)
+            
+            # 转换为units格式并格式化输出（保留真实的文件起始行号）
+            structured_units = []
+            for block in merged_blocks:
+                structured_units.append({
+                    "block_id": block["block_id"],
+                    "content": block["content"],
+                    "start_line": block.get("start_line", 1),
+                })
+            
+            output = self._format_structured_output(filepath, structured_units, total_lines, agent)
+            
+            # 尝试获取上下文信息（使用合并后的范围）
+            all_start_lines = [req.get("start_line", 1) for req in requests]
+            all_end_lines = [req.get("end_line", total_lines) for req in requests]
+            min_start = min(all_start_lines)
+            max_end = max(all_end_lines)
+            context_info = self._get_file_context(filepath, min_start, max_end, agent)
+            if context_info:
+                output += context_info
+            
+            return {"success": True, "stdout": output, "stderr": ""}
+            
+        except Exception as e:
+            return {"success": False, "stdout": "", "stderr": f"合并范围读取失败: {str(e)}"}
+
+    def _get_file_context(
+        self, filepath: str, start_line: int, end_line: int, agent: Any = None
+    ) -> str:
+        """获取文件的上下文信息
+
+        Args:
+            filepath: 文件路径
+            start_line: 起始行号
+            end_line: 结束行号
+            agent: Agent实例
+
+        Returns:
+            格式化的上下文信息字符串，如果无法获取则返回空字符串
+        """
+        try:
+            # 尝试从Agent获取CodeAgent实例
+            if not agent:
+                return ""
+
+            # 通过agent获取CodeAgent实例
+            # CodeAgent在初始化时会将自身关联到agent
+            code_agent = getattr(agent, "_code_agent", None)
+            if not code_agent:
+                return ""
+
+            # 获取上下文管理器
+            context_manager = getattr(code_agent, "context_manager", None)
+            if not context_manager:
+                return ""
+
+            # 输出上下文感知日志
+            file_name = os.path.basename(filepath)
+            if start_line == end_line:
+                line_info = f"第{start_line}行"
+            else:
+                line_info = f"第{start_line}-{end_line}行"
+            print(f"🧠 正在分析代码上下文 ({file_name}, {line_info})...")
+
+            # 确保文件已更新到上下文管理器
+            # 如果文件内容已缓存，直接使用；否则读取并更新
+            if not hasattr(context_manager, "_file_cache") or filepath not in context_manager._file_cache:
+                try:
+                    with open(filepath, "r", encoding="utf-8", errors="replace") as f:
+                        content = f.read()
+                    context_manager.update_context_for_file(filepath, content)
+                except Exception:
+                    # 如果读取失败，尝试获取已有上下文
+                    pass
+
+            # 获取编辑上下文
+            edit_context = context_manager.get_edit_context(filepath, start_line, end_line)
+
+            # 构建上下文信息
+            if not edit_context.context_summary or edit_context.context_summary == "No context available":
+                return ""
+
+            # 格式化上下文信息
+            context_lines = ["\n📋 代码上下文信息:"]
+            context_lines.append("─" * 60)
+
+            if edit_context.current_scope:
+                scope_info = f"📍 当前作用域: {edit_context.current_scope.kind} `{edit_context.current_scope.name}`"
+                if edit_context.current_scope.signature:
+                    scope_info += f"\n   └─ 签名: {edit_context.current_scope.signature}"
+                context_lines.append(scope_info)
+
+            if edit_context.used_symbols:
+                # 对符号去重（基于 name + file_path + line_start）
+                seen_symbols = set()
+                unique_symbols = []
+                for s in edit_context.used_symbols:
+                    key = (s.name, getattr(s, 'file_path', ''), getattr(s, 'line_start', 0))
+                    if key not in seen_symbols:
+                        seen_symbols.add(key)
+                        unique_symbols.append(s)
+                
+                # 区分定义和调用，显示定义位置信息
+                definitions = []
+                calls = []
+                for symbol in unique_symbols[:10]:
+                    is_def = getattr(symbol, 'is_definition', False)
+                    if is_def:
+                        definitions.append(symbol)
+                    else:
+                        calls.append(symbol)
+                
+                # 显示定义
+                if definitions:
+                    def_names = [f"`{s.name}`" for s in definitions]
+                    context_lines.append(f"📝 定义的符号: {', '.join(def_names)}")
+                
+                # 显示调用（带定义位置信息）
+                if calls:
+                    call_info = []
+                    for symbol in calls:
+                        def_loc = getattr(symbol, 'definition_location', None)
+                        if def_loc:
+                            def_file = os.path.basename(def_loc.file_path)
+                            def_line = def_loc.line_start
+                            call_info.append(f"`{symbol.name}` → {def_file}:{def_line}")
+                        else:
+                            call_info.append(f"`{symbol.name}`")
+                    context_lines.append(f"🔗 调用的符号: {', '.join(call_info)}")
+                
+                # 如果还有更多符号
+                more = len(edit_context.used_symbols) - 10
+                if more > 0:
+                    context_lines.append(f"   ... 还有{more}个符号")
+
+            # 不再感知导入符号
+
+            if edit_context.relevant_files:
+                # 对相关文件去重
+                unique_files = list(dict.fromkeys(edit_context.relevant_files))
+                rel_files = unique_files[:10]
+                files_str = "\n   ".join(f"• {os.path.relpath(f, context_manager.project_root)}" for f in rel_files)
+                more = len(unique_files) - 10
+                if more > 0:
+                    files_str += f"\n   ... 还有{more}个相关文件"
+                context_lines.append(f"📁 相关文件 ({len(unique_files)}个):\n   {files_str}")
+
+            context_lines.append("─" * 60)
+            context_lines.append("")  # 空行
+
+            # 打印上下文感知结果到控制台
+            context_output = "\n".join(context_lines)
+            print(f"🧠 上下文感知结果:\n{context_output}")
+            
+            return context_output
+
+        except Exception:
+            # 静默失败，不影响文件读取
+            return ""
+
     def execute(self, args: Dict) -> Dict[str, Any]:
         """执行代码读取操作
 
@@ -146,33 +1276,604 @@ class ReadCodeTool:
                     "stdout": "",
                     "stderr": "参数中必须包含文件列表",
                 }
+            
+            if len(args["files"]) == 0:
+                return {
+                    "success": False,
+                    "stdout": "",
+                    "stderr": "文件列表不能为空",
+                }
 
             all_outputs = []
             overall_success = True
+            status_lines = []
+            total_tokens = 0  # 累计读取的token数
+            max_token_limit = self._get_max_token_limit(agent)
 
+            # 第一遍：检查所有文件的累计token数是否超过限制
+            file_read_info = []  # 存储每个文件要读取的信息
             for file_info in args["files"]:
                 if not isinstance(file_info, dict) or "path" not in file_info:
                     continue
+                
+                filepath = file_info["path"].strip()
+                start_line = file_info.get("start_line", 1)
+                end_line = file_info.get("end_line", -1)
+                
+                # 检查文件是否存在并计算要读取的token数
+                abs_path = os.path.abspath(filepath)
+                if not os.path.exists(abs_path):
+                    continue
+                
+                try:
+                    # 统计总行数
+                    with open(abs_path, "r", encoding="utf-8", errors="ignore") as f:
+                        total_lines = sum(1 for _ in f)
+                    
+                    if total_lines == 0:
+                        continue
+                    
+                    # 计算实际要读取的行范围
+                    if end_line == -1:
+                        actual_end_line = total_lines
+                    else:
+                        actual_end_line = (
+                            max(1, min(end_line, total_lines))
+                            if end_line >= 0
+                            else total_lines + end_line + 1
+                        )
+                    
+                    actual_start_line = (
+                        max(1, min(start_line, total_lines))
+                        if start_line >= 0
+                        else total_lines + start_line + 1
+                    )
+                    
+                    if actual_start_line <= actual_end_line:
+                        # 读取完整文件内容用于token估算
+                        with open(abs_path, "r", encoding="utf-8", errors="ignore") as f:
+                            file_content = f.read()
+                        
+                        # 估算结构化输出的token数
+                        raw_mode = file_info.get("raw_mode", False)
+                        content_tokens = self._estimate_structured_tokens(
+                            abs_path, file_content, actual_start_line, actual_end_line, total_lines, raw_mode
+                        )
+                        
+                        file_read_info.append({
+                            "filepath": filepath,
+                            "start_line": actual_start_line,
+                            "end_line": actual_end_line,
+                            "read_lines": actual_end_line - actual_start_line + 1,
+                            "tokens": content_tokens,
+                            "file_info": file_info,
+                        })
+                        total_tokens += content_tokens
+                except Exception:
+                    continue
 
-                result = self._handle_single_file(
-                    file_info["path"].strip(),
-                    file_info.get("start_line", 1),
-                    file_info.get("end_line", -1),
-                    agent,
+            # 检查累计token数是否超过限制
+            if total_tokens > max_token_limit:
+                file_list = "\n   ".join(
+                    f"• {info['filepath']}: {info['tokens']} tokens ({info['read_lines']} 行, 范围: {info['start_line']}-{info['end_line']})"
+                    for info in file_read_info[:10]
                 )
+                more_files = len(file_read_info) - 10
+                if more_files > 0:
+                    file_list += f"\n   ... 还有 {more_files} 个文件"
+                
+                return {
+                    "success": False,
+                    "stdout": "",
+                    "stderr": (
+                        f"⚠️ 累计读取范围过大: 请求累计读取内容约 {total_tokens} tokens，超过限制 ({max_token_limit} tokens，约2/3最大窗口)\n"
+                        f"📋 文件列表 ({len(file_read_info)} 个文件):\n   {file_list}\n"
+                        f"💡 建议：\n"
+                        f"   1. 分批读取：将文件分成多个批次，每批累计内容不超过 {max_token_limit} tokens\n"
+                        f"   2. 先定位：使用搜索或分析工具定位关键代码位置，再读取具体范围\n"
+                        f"   3. 缩小范围：为每个文件指定更精确的行号范围"
+                    ),
+                }
 
-                if result["success"]:
-                    all_outputs.append(result["stdout"])
+            # 第二遍：实际读取文件（按文件分组，合并同一文件的多个范围请求，避免块重复）
+            # 按文件路径分组
+            from collections import defaultdict
+            file_requests = defaultdict(list)
+            for file_info in args["files"]:
+                if not isinstance(file_info, dict) or "path" not in file_info:
+                    continue
+                abs_path = os.path.abspath(file_info["path"].strip())
+                file_requests[abs_path].append(file_info)
+            
+            # 按文件处理，合并同一文件的多个范围请求
+            for abs_path, requests in file_requests.items():
+                if len(requests) == 1:
+                    # 单个范围请求，直接处理
+                    file_info = requests[0]
+                    result = self._handle_single_file(
+                        file_info["path"].strip(),
+                        file_info.get("start_line", 1),
+                        file_info.get("end_line", -1),
+                        agent,
+                        file_info.get("raw_mode", False),
+                    )
+                    if result["success"]:
+                        all_outputs.append(result["stdout"])
+                        status_lines.append(f"✅ {file_info['path']} 文件读取成功")
+                    else:
+                        all_outputs.append(f"❌ {file_info['path']}: {result['stderr']}")
+                        status_lines.append(f"❌ {file_info['path']} 文件读取失败")
+                        overall_success = False
                 else:
-                    all_outputs.append(f"❌ {file_info['path']}: {result['stderr']}")
-                    overall_success = False
+                    # 多个范围请求，合并处理并去重
+                    merged_result = self._handle_merged_ranges(
+                        abs_path, requests, agent
+                    )
+                    display_path = requests[0]["path"]
+                    if merged_result["success"]:
+                        all_outputs.append(merged_result["stdout"])
+                        status_lines.append(f"✅ {display_path} 文件读取成功 (合并{len(requests)}个范围请求，已去重)")
+                    else:
+                        all_outputs.append(f"❌ {display_path}: {merged_result['stderr']}")
+                        status_lines.append(f"❌ {display_path} 文件读取失败")
+                        overall_success = False
 
+            stdout_text = "\n".join(all_outputs)
+            # 仅打印每个文件的读取状态，不打印具体内容
+            try:
+                if status_lines:
+                    print("\n".join(status_lines), end="\n")
+            except Exception:
+                pass
             return {
                 "success": overall_success,
-                "stdout": "\n".join(all_outputs),
+                "stdout": stdout_text,
                 "stderr": "",
             }
 
         except Exception as e:
-            PrettyOutput.print(str(e), OutputType.ERROR)
+            print(f"❌ {str(e)}")
             return {"success": False, "stdout": "", "stderr": f"代码读取失败: {str(e)}"}
+
+
+def main():
+    """测试结构化读取功能"""
+    import tempfile
+    import os
+    
+    tool = ReadCodeTool()
+    
+    print("=" * 80)
+    print("测试结构化读取功能")
+    print("=" * 80)
+    
+    # 测试1: C语言文件（tree-sitter支持）
+    print("\n【测试1】C语言文件 - 语法单元提取")
+    print("-" * 80)
+    
+    c_code = """#include <stdio.h>
+
+void main() {
+    printf("Hello, World!\\n");
+}
+
+int add(int a, int b) {
+    return a + b;
+}
+
+int sub(int a, int b) {
+    return a - b;
+}
+
+struct Point {
+    int x;
+    int y;
+};
+"""
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.c', delete=False) as f:
+        c_file = f.name
+        f.write(c_code)
+    
+    try:
+        result = tool.execute({
+            "files": [{"path": c_file, "start_line": 1, "end_line": -1}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ C语言文件读取成功")
+            print("\n输出内容:")
+            print(result["stdout"])
+        else:
+            print(f"❌ C语言文件读取失败: {result['stderr']}")
+    finally:
+        os.unlink(c_file)
+    
+    # 测试2: Python文件（AST支持）
+    print("\n【测试2】Python文件 - 语法单元提取")
+    print("-" * 80)
+    
+    python_code = """def main():
+    print("Hello, World!")
+
+def add(a, b):
+    return a + b
+
+def sub(a, b):
+    return a - b
+
+class Point:
+    def __init__(self, x, y):
+        self.x = x
+        self.y = y
+"""
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as f:
+        py_file = f.name
+        f.write(python_code)
+    
+    try:
+        result = tool.execute({
+            "files": [{"path": py_file, "start_line": 1, "end_line": -1}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ Python文件读取成功")
+            print("\n输出内容:")
+            print(result["stdout"])
+        else:
+            print(f"❌ Python文件读取失败: {result['stderr']}")
+    finally:
+        os.unlink(py_file)
+    
+    # 测试3: 不支持的语言 - 行号分组
+    print("\n【测试3】不支持的语言 - 行号分组（20行一组）")
+    print("-" * 80)
+    
+    text_content = "\n".join([f"这是第 {i} 行内容" for i in range(1, 51)])
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.txt', delete=False) as f:
+        txt_file = f.name
+        f.write(text_content)
+    
+    try:
+        result = tool.execute({
+            "files": [{"path": txt_file, "start_line": 1, "end_line": -1}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ 文本文件读取成功（使用行号分组）")
+            print("\n输出内容（前500字符）:")
+            print(result["stdout"][:500] + "..." if len(result["stdout"]) > 500 else result["stdout"])
+        else:
+            print(f"❌ 文本文件读取失败: {result['stderr']}")
+    finally:
+        os.unlink(txt_file)
+    
+    # 测试4: 指定行号范围
+    print("\n【测试4】指定行号范围读取")
+    print("-" * 80)
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.c', delete=False) as f:
+        c_file2 = f.name
+        f.write(c_code)
+    
+    try:
+        result = tool.execute({
+            "files": [{"path": c_file2, "start_line": 1, "end_line": 10}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ 指定范围读取成功")
+            print("\n输出内容:")
+            print(result["stdout"])
+        else:
+            print(f"❌ 指定范围读取失败: {result['stderr']}")
+    finally:
+        os.unlink(c_file2)
+    
+    # 测试5: 边界情况 - 返回边界上的语法单元
+    print("\n【测试5】边界情况 - 返回边界上的语法单元")
+    print("-" * 80)
+    
+    boundary_test_code = """def func1():
+    line1 = 1
+    line2 = 2
+    line3 = 3
+
+def func2():
+    line1 = 1
+    line2 = 2
+
+def func3():
+    line1 = 1
+    line2 = 2
+    line3 = 3
+    line4 = 4
+"""
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as f:
+        boundary_file = f.name
+        f.write(boundary_test_code)
+    
+    try:
+        # 请求第3-8行
+        # func1: 1-4行（结束行4在范围内，应该返回完整func1）
+        # func2: 6-8行（开始行6在范围内，应该返回完整func2）
+        # func3: 10-14行（完全不在范围内，不应该返回）
+        result = tool.execute({
+            "files": [{"path": boundary_file, "start_line": 3, "end_line": 8}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ 边界情况测试成功")
+            print("请求范围: 3-8行")
+            print("预期结果:")
+            print("  - func1 (1-4行): 结束行4在范围内，应返回完整func1")
+            print("  - func2 (6-8行): 开始行6在范围内，应返回完整func2")
+            print("  - func3 (10-14行): 完全不在范围内，不应返回")
+            print("\n实际输出:")
+            print(result["stdout"])
+        else:
+            print(f"❌ 边界情况测试失败: {result['stderr']}")
+    finally:
+        os.unlink(boundary_file)
+    
+    # 测试6: 多个文件
+    print("\n【测试6】多个文件读取")
+    print("-" * 80)
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.c', delete=False) as f1, \
+         tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as f2:
+        c_file3 = f1.name
+        py_file2 = f2.name
+        f1.write(c_code)
+        f2.write(python_code)
+    
+    try:
+        result = tool.execute({
+            "files": [
+                {"path": c_file3, "start_line": 1, "end_line": -1},
+                {"path": py_file2, "start_line": 1, "end_line": -1}
+            ],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ 多文件读取成功")
+            print("\n输出内容（前800字符）:")
+            print(result["stdout"][:800] + "..." if len(result["stdout"]) > 800 else result["stdout"])
+        else:
+            print(f"❌ 多文件读取失败: {result['stderr']}")
+    finally:
+        os.unlink(c_file3)
+        os.unlink(py_file2)
+    
+    # 测试7: 嵌套作用域的边界情况
+    print("\n【测试7】嵌套作用域的边界情况")
+    print("-" * 80)
+    
+    nested_code = """class Outer:
+    def method1(self):
+        line1 = 1
+        line2 = 2
+    
+    def method2(self):
+        line1 = 1
+        line2 = 2
+        line3 = 3
+
+def standalone_func():
+    line1 = 1
+    line2 = 2
+"""
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False) as f:
+        nested_file = f.name
+        f.write(nested_code)
+    
+    try:
+        # 请求第4-7行
+        # Outer.method1: 2-4行（结束行4在范围内，应该返回完整method1）
+        # Outer.method2: 6-9行（开始行6在范围内，应该返回完整method2）
+        # Outer类: 1-9行（包含method1和method2，应该返回）
+        # standalone_func: 11-13行（完全不在范围内，不应返回）
+        result = tool.execute({
+            "files": [{"path": nested_file, "start_line": 4, "end_line": 7}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ 嵌套作用域边界测试成功")
+            print("请求范围: 4-7行")
+            print("预期结果:")
+            print("  - Outer类 (1-9行): 包含method1和method2，应返回")
+            print("  - Outer.method1 (2-4行): 结束行4在范围内，应返回完整method1")
+            print("  - Outer.method2 (6-9行): 开始行6在范围内，应返回完整method2")
+            print("\n实际输出:")
+            print(result["stdout"])
+        else:
+            print(f"❌ 嵌套作用域边界测试失败: {result['stderr']}")
+    finally:
+        os.unlink(nested_file)
+    
+    # 测试8: Java文件（tree-sitter支持）
+    print("\n【测试8】Java文件 - 语法单元提取")
+    print("-" * 80)
+    
+    java_code = """public class Main {
+    public static void main(String[] args) {
+        System.out.println("Hello, World!");
+    }
+    
+    public int add(int a, int b) {
+        return a + b;
+    }
+    
+    private int subtract(int a, int b) {
+        return a - b;
+    }
+}
+
+class Point {
+    private int x;
+    private int y;
+    
+    public Point(int x, int y) {
+        this.x = x;
+        this.y = y;
+    }
+}
+"""
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.java', delete=False) as f:
+        java_file = f.name
+        f.write(java_code)
+    
+    try:
+        result = tool.execute({
+            "files": [{"path": java_file, "start_line": 1, "end_line": -1}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ Java文件读取成功")
+            print("\n输出内容:")
+            print(result["stdout"])
+        else:
+            print(f"❌ Java文件读取失败: {result['stderr']}")
+    finally:
+        os.unlink(java_file)
+    
+    # 测试9: Rust文件（tree-sitter支持）
+    print("\n【测试9】Rust文件 - 语法单元提取")
+    print("-" * 80)
+    
+    rust_code = """fn main() {
+    println!("Hello, World!");
+}
+
+fn add(a: i32, b: i32) -> i32 {
+    a + b
+}
+
+fn subtract(a: i32, b: i32) -> i32 {
+    a - b
+}
+
+struct Point {
+    x: i32,
+    y: i32,
+}
+
+impl Point {
+    fn new(x: i32, y: i32) -> Point {
+        Point { x, y }
+    }
+}
+
+enum Color {
+    Red,
+    Green,
+    Blue,
+}
+"""
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.rs', delete=False) as f:
+        rust_file = f.name
+        f.write(rust_code)
+    
+    try:
+        result = tool.execute({
+            "files": [{"path": rust_file, "start_line": 1, "end_line": -1}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ Rust文件读取成功")
+            print("\n输出内容:")
+            print(result["stdout"])
+        else:
+            print(f"❌ Rust文件读取失败: {result['stderr']}")
+    finally:
+        os.unlink(rust_file)
+    
+    # 测试10: Go文件（tree-sitter支持）
+    print("\n【测试10】Go文件 - 语法单元提取")
+    print("-" * 80)
+    
+    go_code = """package main
+
+import "fmt"
+
+func main() {
+    fmt.Println("Hello, World!")
+}
+
+func add(a int, b int) int {
+    return a + b
+}
+
+func subtract(a int, b int) int {
+    return a - b
+}
+
+type Point struct {
+    x int
+    y int
+}
+
+func (p *Point) New(x int, y int) {
+    p.x = x
+    p.y = y
+}
+
+type Color int
+
+const (
+    Red Color = iota
+    Green
+    Blue
+)
+
+type Shape interface {
+    Area() float64
+    Perimeter() float64
+}
+
+type Drawable interface {
+    Draw()
+}
+"""
+    
+    with tempfile.NamedTemporaryFile(mode='w', suffix='.go', delete=False) as f:
+        go_file = f.name
+        f.write(go_code)
+    
+    try:
+        result = tool.execute({
+            "files": [{"path": go_file, "start_line": 1, "end_line": -1}],
+            "agent": None
+        })
+        
+        if result["success"]:
+            print("✅ Go文件读取成功")
+            print("\n输出内容:")
+            print(result["stdout"])
+        else:
+            print(f"❌ Go文件读取失败: {result['stderr']}")
+    finally:
+        os.unlink(go_file)
+    
+    print("\n" + "=" * 80)
+    print("测试完成")
+    print("=" * 80)
+
+
+if __name__ == "__main__":
+    main()