valkryhx/google_adk_agent / skills/codebase_search
代码库搜索专家
Verified通过 ripgrep 在本地文件系统中精确定位代码、配置或文档,支持多轮迭代检索。
tools.py
View on GitHub"""
Codebase Search Skill - 代码库搜索工具集
基于 ripgrep 的高性能文本检索,用于 Agentic RAG。
"""
import subprocess
import os
from typing import Optional, List
def execute_ripgrep(
pattern: str,
path: str = ".",
context_lines: int = 2,
ignore_case: bool = True,
file_type: Optional[str] = None,
max_results: int = 50
) -> str:
"""
使用 ripgrep 进行高性能文本检索
Args:
pattern: 搜索模式(支持正则表达式)
path: 搜索路径,默认当前目录
context_lines: 显示匹配行的上下文行数
ignore_case: 是否忽略大小写
file_type: 限制文件类型(如 'py', 'md', 'js')
max_results: 最大结果数
Returns:
搜索结果或错误信息
"""
try:
cmd = [
"rg", "--color", "never", "-n",
"-A", str(context_lines),
"-B", str(context_lines),
"-m", str(max_results),
"--iglob", "!nul", # 排除 Windows 保留设备名(大小写不敏感)
"--iglob", "!con", # 排除其他 Windows 保留设备名
"--iglob", "!prn",
"--iglob", "!aux",
]
if ignore_case:
cmd.append("-i")
if file_type:
cmd.extend(["-t", file_type])
cmd.extend([pattern, path])
result = subprocess.run(
cmd,
capture_output=True,
text=True,
encoding='utf-8',
timeout=30
)
if result.returncode == 0:
output = result.stdout
# 限制输出长度
if len(output) > 5000:
lines = output.split('\n')
output = '\n'.join(lines[:100])
output += f"\n\n... [结果已截断,共匹配更多内容,请使用更精确的搜索条件]"
return output if output else "匹配成功但无输出。"
elif result.returncode == 1:
return "[WARN] 未找到匹配内容。建议:尝试简化关键词或使用通配符如 `.*`"
else:
return f"[ERROR] Ripgrep 错误: {result.stderr}"
except FileNotFoundError:
return "[ERROR] 系统未安装 ripgrep (rg)。请先安装:\n Windows: choco install ripgrep\n macOS: brew install ripgrep\n Linux: apt install ripgrep"
except subprocess.TimeoutExpired:
return "[ERROR] 搜索超时,请缩小搜索范围或简化正则表达式。"
except Exception as e:
return f"[ERROR] 执行异常: {type(e).__name__}: {str(e)}"
def read_file_content(
file_path: str,
start_line: int = 1,
end_line: Optional[int] = None,
max_chars: int = 10000
) -> str:
"""
读取文件内容
Args:
file_path: 文件路径
start_line: 起始行号(1-indexed)
end_line: 结束行号(可选,不指定则读到文件末尾)
max_chars: 最大字符数
Returns:
文件内容或错误信息
"""
try:
with open(file_path, 'r', encoding='utf-8') as f:
lines = f.readlines()
total_lines = len(lines)
# 处理行号范围
start_idx = max(0, start_line - 1)
end_idx = end_line if end_line else total_lines
selected_lines = lines[start_idx:end_idx]
# 添加行号
numbered_content = []
for i, line in enumerate(selected_lines, start=start_line):
numbered_content.append(f"{i:4d}: {line.rstrip()}")
content = '\n'.join(numbered_content)
# 限制长度
if len(content) > max_chars:
content = content[:max_chars] + f"\n\n... [内容已截断,文件共 {total_lines} 行]"
header = f"[文件] {file_path} (第 {start_line}-{min(end_idx, total_lines)} 行,共 {total_lines} 行)\n"
return header + "=" * 50 + "\n" + content
except FileNotFoundError:
return f"[ERROR] 文件不存在: {file_path}"
except UnicodeDecodeError:
return f"[ERROR] 无法读取二进制文件: {file_path}"
except Exception as e:
return f"[ERROR] 读取失败: {type(e).__name__}: {str(e)}"
def list_files(
path: str = ".",
pattern: str = "*",
max_depth: int = 3,
file_type: Optional[str] = None
) -> str:
"""
列出目录下的文件
Args:
path: 目录路径
pattern: 文件名 glob 模式
max_depth: 最大搜索深度
file_type: 文件类型过滤
Returns:
文件列表
"""
try:
cmd = [
"rg", "--files", "--max-depth", str(max_depth),
"--iglob", "!nul", # 排除 Windows 保留设备名
"--iglob", "!con",
"--iglob", "!prn",
"--iglob", "!aux",
]
if pattern != "*":
cmd.extend(["--glob", pattern])
if file_type:
cmd.extend(["-t", file_type])
cmd.append(path)
result = subprocess.run(
cmd,
capture_output=True,
text=True,
encoding='utf-8',
timeout=10
)
if result.returncode == 0 and result.stdout:
files = result.stdout.strip().split('\n')
output = f"📁 目录: {path} (找到 {len(files)} 个文件)\n"
output += "=" * 50 + "\n"
if len(files) > 50:
output += '\n'.join(files[:50])
output += f"\n\n... [仅显示前 50 个,共 {len(files)} 个文件]"
else:
output += '\n'.join(files)
return output
else:
return f"[WARN] 未找到匹配文件"
except FileNotFoundError:
# 如果 rg 不可用,使用 os.walk 作为备选
try:
files = []
for root, dirs, filenames in os.walk(path):
depth = root.replace(path, '').count(os.sep)
if depth >= max_depth:
dirs[:] = []
continue
for f in filenames:
files.append(os.path.join(root, f))
output = f"📁 目录: {path} (找到 {len(files)} 个文件)\n"
output += "=" * 50 + "\n"
output += '\n'.join(files[:50])
if len(files) > 50:
output += f"\n\n... [仅显示前 50 个]"
return output
except Exception as e:
return f"[ERROR] 列出文件失败: {str(e)}"
except Exception as e:
return f"[ERROR] 列出文件失败: {type(e).__name__}: {str(e)}"
def search_and_read(
pattern: str,
path: str = ".",
read_first_match: bool = True
) -> str:
"""
搜索并自动读取第一个匹配文件的内容(便捷方法)
Args:
pattern: 搜索模式
path: 搜索路径
read_first_match: 是否自动读取第一个匹配的文件
Returns:
搜索结果和文件内容
"""
# 先执行搜索
search_result = execute_ripgrep(pattern, path, context_lines=0, max_results=10)
if "未找到" in search_result or "错误" in search_result:
return search_result
output = ["[搜索结果]", search_result]
if read_first_match:
# 提取第一个文件路径
lines = search_result.strip().split('\n')
if lines:
first_line = lines[0]
if ':' in first_line:
file_path = first_line.split(':')[0]
output.append("\n" + "=" * 50)
output.append(f"\n[读取] 自动读取文件: {file_path}\n")
file_content = read_file_content(file_path)
output.append(file_content)
return '\n'.join(output)
# 工具函数字典
CODEBASE_SEARCH_TOOLS = {
"execute_ripgrep": execute_ripgrep,
"read_file_content": read_file_content,
"list_files": list_files,
"search_and_read": search_and_read,
}
def get_tools(*args, **kwargs) -> List:
"""返回所有代码搜索工具函数列表"""
return list(CODEBASE_SEARCH_TOOLS.values())
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |