mirror of
https://github.com/DrizzleTime/Foxel.git
synced 2026-05-11 09:59:50 +08:00
feat: add tools for web fetching
This commit is contained in:
shiyu
@@ -32,6 +32,7 @@ def _build_system_prompt(current_path: Optional[str]) -> str:
|
||||
"",
|
||||
"可用工具：",
|
||||
"- time：获取服务器当前时间（精确到秒，英文星期），支持 year/month/day/hour/minute/second 偏移。",
|
||||
"- web_fetch：抓取网页（HTTP 请求），支持 GET/POST/PUT/PATCH/DELETE/HEAD/OPTIONS，返回状态/标题/正文/链接等。",
|
||||
"- vfs_list_dir：浏览目录（列出 entries + pagination）。",
|
||||
"- vfs_stat：查看文件/目录信息。",
|
||||
"- vfs_read_text：读取文本文件内容（不支持二进制）。",
|
||||
@@ -46,7 +47,7 @@ def _build_system_prompt(current_path: Optional[str]) -> str:
|
||||
"- processors_run：运行处理器处理文件或目录（会返回 task_id 或 task_ids）。",
|
||||
"",
|
||||
"规则：",
|
||||
"1) 读操作（vfs_list_dir/vfs_stat/vfs_read_text/vfs_search）可直接调用工具。",
|
||||
"1) 读操作（web_fetch/vfs_list_dir/vfs_stat/vfs_read_text/vfs_search）可直接调用工具。",
|
||||
"2) 写/改/删操作（vfs_write_text/vfs_mkdir/vfs_delete/vfs_move/vfs_copy/vfs_rename/processors_run）默认需要用户确认；只有在开启自动执行时才应直接执行。",
|
||||
"3) 用户未给出明确路径时先追问；若提供了“当前文件管理目录”，可以基于它把相对描述补全为绝对路径（以 / 开头）。",
|
||||
"4) 修改文件内容：先读取（vfs_read_text）→给出改动点→确认后再写入（vfs_write_text）。",
|
||||
|
||||
@@ -1,623 +0,0 @@
|
||||
import calendar
|
||||
import json
|
||||
from dataclasses import dataclass
|
||||
from datetime import datetime, timedelta
|
||||
from typing import Any, Awaitable, Callable, Dict, List, Optional
|
||||
|
||||
from domain.processors import ProcessDirectoryRequest, ProcessRequest, ProcessorService
|
||||
from domain.virtual_fs import VirtualFSService
|
||||
from domain.virtual_fs.search import VirtualFSSearchService
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class ToolSpec:
|
||||
name: str
|
||||
description: str
|
||||
parameters: Dict[str, Any]
|
||||
requires_confirmation: bool
|
||||
handler: Callable[[Dict[str, Any]], Awaitable[Any]]
|
||||
|
||||
|
||||
def _parse_offset(args: Dict[str, Any], key: str) -> int:
|
||||
value = args.get(key)
|
||||
if value is None:
|
||||
return 0
|
||||
try:
|
||||
return int(value)
|
||||
except (TypeError, ValueError):
|
||||
return 0
|
||||
|
||||
|
||||
def _add_months(dt: datetime, months: int) -> datetime:
|
||||
if months == 0:
|
||||
return dt
|
||||
total = dt.year * 12 + (dt.month - 1) + months
|
||||
year = total // 12
|
||||
month = total % 12 + 1
|
||||
last_day = calendar.monthrange(year, month)[1]
|
||||
day = min(dt.day, last_day)
|
||||
return dt.replace(year=year, month=month, day=day)
|
||||
|
||||
|
||||
async def _time(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
now = datetime.now()
|
||||
year_offset = _parse_offset(args, "year")
|
||||
month_offset = _parse_offset(args, "month")
|
||||
day_offset = _parse_offset(args, "day")
|
||||
hour_offset = _parse_offset(args, "hour")
|
||||
minute_offset = _parse_offset(args, "minute")
|
||||
second_offset = _parse_offset(args, "second")
|
||||
|
||||
dt = _add_months(now, year_offset * 12 + month_offset)
|
||||
dt = dt + timedelta(days=day_offset, hours=hour_offset, minutes=minute_offset, seconds=second_offset)
|
||||
|
||||
weekday_names = [
|
||||
"Monday",
|
||||
"Tuesday",
|
||||
"Wednesday",
|
||||
"Thursday",
|
||||
"Friday",
|
||||
"Saturday",
|
||||
"Sunday",
|
||||
]
|
||||
weekday = weekday_names[dt.weekday()]
|
||||
dt_str = dt.strftime("%Y-%m-%d %H:%M:%S")
|
||||
return {
|
||||
"ok": True,
|
||||
"summary": f"{dt_str} · {weekday}",
|
||||
"data": {
|
||||
"datetime": dt_str,
|
||||
"weekday": weekday,
|
||||
"offset": {
|
||||
"year": year_offset,
|
||||
"month": month_offset,
|
||||
"day": day_offset,
|
||||
"hour": hour_offset,
|
||||
"minute": minute_offset,
|
||||
"second": second_offset,
|
||||
},
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
async def _processors_list(_: Dict[str, Any]) -> Dict[str, Any]:
|
||||
return {"processors": ProcessorService.list_processors()}
|
||||
|
||||
|
||||
async def _processors_run(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = str(args.get("path") or "")
|
||||
processor_type = str(args.get("processor_type") or "")
|
||||
config = args.get("config")
|
||||
if not isinstance(config, dict):
|
||||
config = {}
|
||||
|
||||
save_to = args.get("save_to")
|
||||
save_to = str(save_to) if isinstance(save_to, str) and save_to.strip() else None
|
||||
|
||||
max_depth = args.get("max_depth")
|
||||
max_depth_value: Optional[int] = None
|
||||
if max_depth is not None:
|
||||
try:
|
||||
max_depth_value = int(max_depth)
|
||||
except (TypeError, ValueError):
|
||||
max_depth_value = None
|
||||
|
||||
suffix = args.get("suffix")
|
||||
suffix_value = str(suffix) if isinstance(suffix, str) and suffix.strip() else None
|
||||
|
||||
overwrite_value = args.get("overwrite")
|
||||
overwrite = bool(overwrite_value) if overwrite_value is not None else None
|
||||
|
||||
is_dir = await VirtualFSService.path_is_directory(path)
|
||||
if is_dir and (max_depth_value is not None or suffix_value is not None):
|
||||
req = ProcessDirectoryRequest(
|
||||
path=path,
|
||||
processor_type=processor_type,
|
||||
config=config,
|
||||
overwrite=True if overwrite is None else overwrite,
|
||||
max_depth=max_depth_value,
|
||||
suffix=suffix_value,
|
||||
)
|
||||
result = await ProcessorService.process_directory(req)
|
||||
return {"mode": "directory", **result}
|
||||
|
||||
req = ProcessRequest(
|
||||
path=path,
|
||||
processor_type=processor_type,
|
||||
config=config,
|
||||
save_to=save_to,
|
||||
overwrite=False if overwrite is None else overwrite,
|
||||
)
|
||||
result = await ProcessorService.process_file(req)
|
||||
return {"mode": "file", **result}
|
||||
|
||||
|
||||
def _normalize_vfs_path(value: Any) -> str:
|
||||
s = str(value or "").strip().replace("\\", "/")
|
||||
if not s:
|
||||
return ""
|
||||
if not s.startswith("/"):
|
||||
s = "/" + s
|
||||
s = s.rstrip("/") or "/"
|
||||
return s
|
||||
|
||||
|
||||
def _require_vfs_path(value: Any, field: str) -> str:
|
||||
path = _normalize_vfs_path(value)
|
||||
if not path:
|
||||
raise ValueError(f"missing_{field}")
|
||||
return path
|
||||
|
||||
|
||||
async def _vfs_list_dir(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _normalize_vfs_path(args.get("path") or "/") or "/"
|
||||
page = int(args.get("page") or 1)
|
||||
page_size = int(args.get("page_size") or 50)
|
||||
sort_by = str(args.get("sort_by") or "name")
|
||||
sort_order = str(args.get("sort_order") or "asc")
|
||||
return await VirtualFSService.list_directory(path, page, page_size, sort_by, sort_order)
|
||||
|
||||
|
||||
async def _vfs_stat(args: Dict[str, Any]) -> Any:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
return await VirtualFSService.stat(path)
|
||||
|
||||
|
||||
async def _vfs_read_text(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
encoding = str(args.get("encoding") or "utf-8")
|
||||
max_chars = int(args.get("max_chars") or 8000)
|
||||
|
||||
data = await VirtualFSService.read_file(path)
|
||||
if isinstance(data, (bytes, bytearray)):
|
||||
try:
|
||||
text = bytes(data).decode(encoding)
|
||||
except UnicodeDecodeError:
|
||||
return {"error": "binary_or_invalid_text", "path": path}
|
||||
elif isinstance(data, str):
|
||||
text = data
|
||||
else:
|
||||
text = str(data)
|
||||
|
||||
original_len = len(text)
|
||||
truncated = original_len > max_chars
|
||||
if truncated:
|
||||
text = text[:max_chars]
|
||||
return {
|
||||
"path": path,
|
||||
"encoding": encoding,
|
||||
"content": text,
|
||||
"truncated": truncated,
|
||||
"length": original_len,
|
||||
}
|
||||
|
||||
|
||||
async def _vfs_write_text(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
if path == "/":
|
||||
raise ValueError("invalid_path")
|
||||
encoding = str(args.get("encoding") or "utf-8")
|
||||
content = str(args.get("content") or "")
|
||||
data = content.encode(encoding)
|
||||
await VirtualFSService.write_file(path, data)
|
||||
return {"written": True, "path": path, "encoding": encoding, "bytes": len(data)}
|
||||
|
||||
|
||||
async def _vfs_mkdir(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
return await VirtualFSService.mkdir(path)
|
||||
|
||||
|
||||
async def _vfs_delete(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
return await VirtualFSService.delete(path)
|
||||
|
||||
|
||||
async def _vfs_move(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
src = _require_vfs_path(args.get("src"), "src")
|
||||
dst = _require_vfs_path(args.get("dst"), "dst")
|
||||
if src == "/" or dst == "/":
|
||||
raise ValueError("invalid_path")
|
||||
overwrite = bool(args.get("overwrite") or False)
|
||||
return await VirtualFSService.move(src, dst, overwrite)
|
||||
|
||||
|
||||
async def _vfs_copy(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
src = _require_vfs_path(args.get("src"), "src")
|
||||
dst = _require_vfs_path(args.get("dst"), "dst")
|
||||
if src == "/" or dst == "/":
|
||||
raise ValueError("invalid_path")
|
||||
overwrite = bool(args.get("overwrite") or False)
|
||||
return await VirtualFSService.copy(src, dst, overwrite)
|
||||
|
||||
|
||||
async def _vfs_rename(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
src = _require_vfs_path(args.get("src"), "src")
|
||||
dst = _require_vfs_path(args.get("dst"), "dst")
|
||||
if src == "/" or dst == "/":
|
||||
raise ValueError("invalid_path")
|
||||
overwrite = bool(args.get("overwrite") or False)
|
||||
return await VirtualFSService.rename(src, dst, overwrite)
|
||||
|
||||
|
||||
async def _vfs_search(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
q = str(args.get("q") or "").strip()
|
||||
if not q:
|
||||
raise ValueError("missing_q")
|
||||
mode = str(args.get("mode") or "vector")
|
||||
top_k = int(args.get("top_k") or 10)
|
||||
page = int(args.get("page") or 1)
|
||||
page_size = int(args.get("page_size") or 10)
|
||||
return await VirtualFSSearchService.search(q, top_k, mode, page, page_size)
|
||||
|
||||
|
||||
TOOLS: Dict[str, ToolSpec] = {
|
||||
"time": ToolSpec(
|
||||
name="time",
|
||||
description=(
|
||||
"获取服务器当前时间（精确到秒，含英文星期）。"
|
||||
" 支持 year/month/day/hour/minute/second 偏移（可为负数）。"
|
||||
),
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"year": {"type": "integer", "description": "年偏移（可为负数）"},
|
||||
"month": {"type": "integer", "description": "月偏移（可为负数）"},
|
||||
"day": {"type": "integer", "description": "日偏移（可为负数）"},
|
||||
"hour": {"type": "integer", "description": "时偏移（可为负数）"},
|
||||
"minute": {"type": "integer", "description": "分偏移（可为负数）"},
|
||||
"second": {"type": "integer", "description": "秒偏移（可为负数）"},
|
||||
},
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_time,
|
||||
),
|
||||
"processors_list": ToolSpec(
|
||||
name="processors_list",
|
||||
description="获取可用处理器列表（type/name/config_schema 等）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {},
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_processors_list,
|
||||
),
|
||||
"processors_run": ToolSpec(
|
||||
name="processors_run",
|
||||
description=(
|
||||
"运行处理器处理文件或目录。"
|
||||
" 对目录可选 max_depth/suffix；对文件可选 overwrite/save_to。"
|
||||
" 返回任务 id（去任务队列查看进度）。"
|
||||
),
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "文件或目录路径（绝对路径，如 /foo/bar）"},
|
||||
"processor_type": {"type": "string", "description": "处理器类型（例如 image_watermark）"},
|
||||
"config": {"type": "object", "description": "处理器配置，按 processors_list 返回的 config_schema 填写"},
|
||||
"overwrite": {"type": "boolean", "description": "是否覆盖原文件/目录内文件"},
|
||||
"save_to": {"type": "string", "description": "保存到指定路径（仅文件模式，且 overwrite=false 时使用）"},
|
||||
"max_depth": {"type": "integer", "description": "目录遍历深度（仅目录模式）"},
|
||||
"suffix": {"type": "string", "description": "目录批处理时的输出后缀（仅 produces_file 且 overwrite=false）"},
|
||||
},
|
||||
"required": ["path", "processor_type"],
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_processors_run,
|
||||
),
|
||||
"vfs_list_dir": ToolSpec(
|
||||
name="vfs_list_dir",
|
||||
description="浏览目录（列出 entries + pagination）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "目录路径（绝对路径，如 /foo/bar）"},
|
||||
"page": {"type": "integer", "description": "页码（从 1 开始）"},
|
||||
"page_size": {"type": "integer", "description": "每页条数"},
|
||||
"sort_by": {"type": "string", "description": "排序字段：name/size/mtime"},
|
||||
"sort_order": {"type": "string", "description": "排序顺序：asc/desc"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_list_dir,
|
||||
),
|
||||
"vfs_stat": ToolSpec(
|
||||
name="vfs_stat",
|
||||
description="查看文件/目录信息（size/mtime/is_dir/has_thumbnail/vector_index 等）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "路径（绝对路径，如 /foo/bar.txt）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_stat,
|
||||
),
|
||||
"vfs_read_text": ToolSpec(
|
||||
name="vfs_read_text",
|
||||
description="读取文本文件内容（解码失败视为二进制，返回 error）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "文件路径（绝对路径，如 /foo/bar.md）"},
|
||||
"encoding": {"type": "string", "description": "文本编码（默认 utf-8）"},
|
||||
"max_chars": {"type": "integer", "description": "最多返回的字符数（默认 8000）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_read_text,
|
||||
),
|
||||
"vfs_write_text": ToolSpec(
|
||||
name="vfs_write_text",
|
||||
description="写入文本文件内容（会覆盖目标文件）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "文件路径（绝对路径，如 /foo/bar.md）"},
|
||||
"content": {"type": "string", "description": "要写入的文本内容"},
|
||||
"encoding": {"type": "string", "description": "文本编码（默认 utf-8）"},
|
||||
},
|
||||
"required": ["path", "content"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_write_text,
|
||||
),
|
||||
"vfs_mkdir": ToolSpec(
|
||||
name="vfs_mkdir",
|
||||
description="创建目录。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "目录路径（绝对路径，如 /foo/bar）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_mkdir,
|
||||
),
|
||||
"vfs_delete": ToolSpec(
|
||||
name="vfs_delete",
|
||||
description="删除文件或目录（由底层适配器决定是否递归）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "路径（绝对路径，如 /foo/bar 或 /foo/bar.txt）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_delete,
|
||||
),
|
||||
"vfs_move": ToolSpec(
|
||||
name="vfs_move",
|
||||
description="移动路径（可能进入任务队列）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"src": {"type": "string", "description": "源路径（绝对路径）"},
|
||||
"dst": {"type": "string", "description": "目标路径（绝对路径）"},
|
||||
"overwrite": {"type": "boolean", "description": "是否允许覆盖已存在目标（默认 false）"},
|
||||
},
|
||||
"required": ["src", "dst"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_move,
|
||||
),
|
||||
"vfs_copy": ToolSpec(
|
||||
name="vfs_copy",
|
||||
description="复制路径（可能进入任务队列）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"src": {"type": "string", "description": "源路径（绝对路径）"},
|
||||
"dst": {"type": "string", "description": "目标路径（绝对路径）"},
|
||||
"overwrite": {"type": "boolean", "description": "是否覆盖已存在目标（默认 false）"},
|
||||
},
|
||||
"required": ["src", "dst"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_copy,
|
||||
),
|
||||
"vfs_rename": ToolSpec(
|
||||
name="vfs_rename",
|
||||
description="重命名路径（本质是同目录 move）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"src": {"type": "string", "description": "源路径（绝对路径）"},
|
||||
"dst": {"type": "string", "description": "目标路径（绝对路径）"},
|
||||
"overwrite": {"type": "boolean", "description": "是否允许覆盖已存在目标（默认 false）"},
|
||||
},
|
||||
"required": ["src", "dst"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_rename,
|
||||
),
|
||||
"vfs_search": ToolSpec(
|
||||
name="vfs_search",
|
||||
description="搜索文件（mode=vector 或 filename）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"q": {"type": "string", "description": "搜索关键词"},
|
||||
"mode": {"type": "string", "description": "搜索模式：vector/filename（默认 vector）"},
|
||||
"top_k": {"type": "integer", "description": "返回数量（vector 模式使用，默认 10）"},
|
||||
"page": {"type": "integer", "description": "页码（filename 模式使用，默认 1）"},
|
||||
"page_size": {"type": "integer", "description": "分页大小（filename 模式使用，默认 10）"},
|
||||
},
|
||||
"required": ["q"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_search,
|
||||
),
|
||||
}
|
||||
|
||||
|
||||
def get_tool(name: str) -> Optional[ToolSpec]:
|
||||
return TOOLS.get(name)
|
||||
|
||||
|
||||
def openai_tools() -> List[Dict[str, Any]]:
|
||||
out: List[Dict[str, Any]] = []
|
||||
for spec in TOOLS.values():
|
||||
out.append({
|
||||
"type": "function",
|
||||
"function": {
|
||||
"name": spec.name,
|
||||
"description": spec.description,
|
||||
"parameters": spec.parameters,
|
||||
},
|
||||
})
|
||||
return out
|
||||
|
||||
|
||||
def _stringify_value(value: Any) -> str:
|
||||
if value is None:
|
||||
return ""
|
||||
if isinstance(value, bool):
|
||||
return "true" if value else "false"
|
||||
if isinstance(value, (int, float)):
|
||||
return str(value)
|
||||
if isinstance(value, str):
|
||||
return value
|
||||
try:
|
||||
return json.dumps(value, ensure_ascii=False)
|
||||
except TypeError:
|
||||
return str(value)
|
||||
|
||||
|
||||
def _list_to_view_items(items: List[Any]) -> List[Any]:
|
||||
normalized: List[Any] = []
|
||||
for item in items:
|
||||
if isinstance(item, dict):
|
||||
normalized.append({str(k): _stringify_value(v) for k, v in item.items()})
|
||||
else:
|
||||
normalized.append(_stringify_value(item))
|
||||
return normalized
|
||||
|
||||
|
||||
def _dict_to_kv_items(data: Dict[str, Any]) -> List[Dict[str, str]]:
|
||||
return [{"key": str(k), "value": _stringify_value(v)} for k, v in data.items()]
|
||||
|
||||
|
||||
def _first_list_field(data: Dict[str, Any]) -> tuple[Optional[str], Optional[List[Any]]]:
|
||||
for key, value in data.items():
|
||||
if isinstance(value, list):
|
||||
return str(key), value
|
||||
return None, None
|
||||
|
||||
|
||||
def _build_view(data: Any) -> Dict[str, Any]:
|
||||
if data is None:
|
||||
return {"type": "kv", "items": []}
|
||||
if isinstance(data, str):
|
||||
return {"type": "text", "text": data}
|
||||
if isinstance(data, list):
|
||||
return {"type": "list", "items": _list_to_view_items(data)}
|
||||
if isinstance(data, dict):
|
||||
content = data.get("content")
|
||||
if isinstance(content, str):
|
||||
meta = {k: _stringify_value(v) for k, v in data.items() if k != "content"}
|
||||
view: Dict[str, Any] = {"type": "text", "text": content}
|
||||
if meta:
|
||||
view["meta"] = meta
|
||||
return view
|
||||
list_key, list_val = _first_list_field(data)
|
||||
if list_key and isinstance(list_val, list):
|
||||
meta = {k: _stringify_value(v) for k, v in data.items() if k != list_key}
|
||||
view = {"type": "list", "title": list_key, "items": _list_to_view_items(list_val)}
|
||||
if meta:
|
||||
view["meta"] = meta
|
||||
return view
|
||||
return {"type": "kv", "items": _dict_to_kv_items(data)}
|
||||
return {"type": "text", "text": _stringify_value(data)}
|
||||
|
||||
|
||||
def _build_summary(view: Dict[str, Any]) -> str:
|
||||
view_type = str(view.get("type") or "")
|
||||
if view_type == "text":
|
||||
text = view.get("text")
|
||||
size = len(text) if isinstance(text, str) else 0
|
||||
return f"chars: {size}" if size else "text"
|
||||
if view_type == "list":
|
||||
items = view.get("items")
|
||||
count = len(items) if isinstance(items, list) else 0
|
||||
title = str(view.get("title") or "items")
|
||||
return f"{title}: {count}"
|
||||
if view_type == "kv":
|
||||
items = view.get("items")
|
||||
count = len(items) if isinstance(items, list) else 0
|
||||
return f"fields: {count}"
|
||||
if view_type == "error":
|
||||
return str(view.get("message") or "error")
|
||||
return ""
|
||||
|
||||
|
||||
def _build_error_payload(code: str, message: str, detail: Any = None) -> Dict[str, Any]:
|
||||
summary = "Canceled" if code == "canceled" else message or "error"
|
||||
view = {"type": "error", "message": summary}
|
||||
payload: Dict[str, Any] = {
|
||||
"ok": False,
|
||||
"summary": summary,
|
||||
"view": view,
|
||||
"error": {
|
||||
"code": code,
|
||||
"message": message,
|
||||
},
|
||||
}
|
||||
if detail is not None:
|
||||
payload["error"]["detail"] = detail
|
||||
return payload
|
||||
|
||||
|
||||
def _normalize_tool_result(result: Any) -> Dict[str, Any]:
|
||||
if isinstance(result, dict) and "ok" in result:
|
||||
payload = dict(result)
|
||||
if payload.get("ok") is False:
|
||||
error = payload.get("error")
|
||||
message = _stringify_value(error.get("message") if isinstance(error, dict) else error)
|
||||
payload.setdefault("summary", message or "error")
|
||||
payload.setdefault("view", {"type": "error", "message": payload["summary"]})
|
||||
return payload
|
||||
data = payload.get("data")
|
||||
if payload.get("view") is None:
|
||||
payload["view"] = _build_view(data)
|
||||
if not payload.get("summary"):
|
||||
payload["summary"] = _build_summary(payload["view"])
|
||||
return payload
|
||||
|
||||
if isinstance(result, dict) and result.get("canceled"):
|
||||
reason = _stringify_value(result.get("reason") or "canceled")
|
||||
return _build_error_payload("canceled", reason, detail=result)
|
||||
|
||||
if isinstance(result, dict) and "error" in result:
|
||||
error = result.get("error")
|
||||
message = _stringify_value(error.get("message") if isinstance(error, dict) else error)
|
||||
return _build_error_payload("error", message, detail=error)
|
||||
|
||||
view = _build_view(result)
|
||||
summary = _build_summary(view)
|
||||
return {"ok": True, "summary": summary, "view": view, "data": result}
|
||||
|
||||
|
||||
def tool_result_to_content(result: Any) -> str:
|
||||
payload = _normalize_tool_result(result)
|
||||
try:
|
||||
return json.dumps(payload, ensure_ascii=False, default=str)
|
||||
except TypeError:
|
||||
return json.dumps({"ok": False, "summary": "error", "view": {"type": "error", "message": "error"}}, ensure_ascii=False)
|
||||
37
domain/agent/tools/__init__.py
Normal file
37
domain/agent/tools/__init__.py
Normal file
@@ -0,0 +1,37 @@
|
||||
from typing import Any, Dict, List, Optional
|
||||
|
||||
from .base import ToolSpec, tool_result_to_content
|
||||
from .processors import TOOLS as PROCESSOR_TOOLS
|
||||
from .time import TOOLS as TIME_TOOLS
|
||||
from .vfs import TOOLS as VFS_TOOLS
|
||||
from .web_fetch import TOOLS as WEB_FETCH_TOOLS
|
||||
|
||||
TOOLS: Dict[str, ToolSpec] = {}
|
||||
for group in (TIME_TOOLS, WEB_FETCH_TOOLS, PROCESSOR_TOOLS, VFS_TOOLS):
|
||||
TOOLS.update(group)
|
||||
|
||||
|
||||
def get_tool(name: str) -> Optional[ToolSpec]:
|
||||
return TOOLS.get(name)
|
||||
|
||||
|
||||
def openai_tools() -> List[Dict[str, Any]]:
|
||||
out: List[Dict[str, Any]] = []
|
||||
for spec in TOOLS.values():
|
||||
out.append({
|
||||
"type": "function",
|
||||
"function": {
|
||||
"name": spec.name,
|
||||
"description": spec.description,
|
||||
"parameters": spec.parameters,
|
||||
},
|
||||
})
|
||||
return out
|
||||
|
||||
|
||||
__all__ = [
|
||||
"ToolSpec",
|
||||
"get_tool",
|
||||
"openai_tools",
|
||||
"tool_result_to_content",
|
||||
]
|
||||
149
domain/agent/tools/base.py
Normal file
149
domain/agent/tools/base.py
Normal file
@@ -0,0 +1,149 @@
|
||||
import json
|
||||
from dataclasses import dataclass
|
||||
from typing import Any, Awaitable, Callable, Dict, List, Optional
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class ToolSpec:
|
||||
name: str
|
||||
description: str
|
||||
parameters: Dict[str, Any]
|
||||
requires_confirmation: bool
|
||||
handler: Callable[[Dict[str, Any]], Awaitable[Any]]
|
||||
|
||||
|
||||
def _stringify_value(value: Any) -> str:
|
||||
if value is None:
|
||||
return ""
|
||||
if isinstance(value, bool):
|
||||
return "true" if value else "false"
|
||||
if isinstance(value, (int, float)):
|
||||
return str(value)
|
||||
if isinstance(value, str):
|
||||
return value
|
||||
try:
|
||||
return json.dumps(value, ensure_ascii=False)
|
||||
except TypeError:
|
||||
return str(value)
|
||||
|
||||
|
||||
def _list_to_view_items(items: List[Any]) -> List[Any]:
|
||||
normalized: List[Any] = []
|
||||
for item in items:
|
||||
if isinstance(item, dict):
|
||||
normalized.append({str(k): _stringify_value(v) for k, v in item.items()})
|
||||
else:
|
||||
normalized.append(_stringify_value(item))
|
||||
return normalized
|
||||
|
||||
|
||||
def _dict_to_kv_items(data: Dict[str, Any]) -> List[Dict[str, str]]:
|
||||
return [{"key": str(k), "value": _stringify_value(v)} for k, v in data.items()]
|
||||
|
||||
|
||||
def _first_list_field(data: Dict[str, Any]) -> tuple[Optional[str], Optional[List[Any]]]:
|
||||
for key, value in data.items():
|
||||
if isinstance(value, list):
|
||||
return str(key), value
|
||||
return None, None
|
||||
|
||||
|
||||
def _build_view(data: Any) -> Dict[str, Any]:
|
||||
if data is None:
|
||||
return {"type": "kv", "items": []}
|
||||
if isinstance(data, str):
|
||||
return {"type": "text", "text": data}
|
||||
if isinstance(data, list):
|
||||
return {"type": "list", "items": _list_to_view_items(data)}
|
||||
if isinstance(data, dict):
|
||||
content = data.get("content")
|
||||
if isinstance(content, str):
|
||||
meta = {k: _stringify_value(v) for k, v in data.items() if k != "content"}
|
||||
view: Dict[str, Any] = {"type": "text", "text": content}
|
||||
if meta:
|
||||
view["meta"] = meta
|
||||
return view
|
||||
list_key, list_val = _first_list_field(data)
|
||||
if list_key and isinstance(list_val, list):
|
||||
meta = {k: _stringify_value(v) for k, v in data.items() if k != list_key}
|
||||
view = {"type": "list", "title": list_key, "items": _list_to_view_items(list_val)}
|
||||
if meta:
|
||||
view["meta"] = meta
|
||||
return view
|
||||
return {"type": "kv", "items": _dict_to_kv_items(data)}
|
||||
return {"type": "text", "text": _stringify_value(data)}
|
||||
|
||||
|
||||
def _build_summary(view: Dict[str, Any]) -> str:
|
||||
view_type = str(view.get("type") or "")
|
||||
if view_type == "text":
|
||||
text = view.get("text")
|
||||
size = len(text) if isinstance(text, str) else 0
|
||||
return f"chars: {size}" if size else "text"
|
||||
if view_type == "list":
|
||||
items = view.get("items")
|
||||
count = len(items) if isinstance(items, list) else 0
|
||||
title = str(view.get("title") or "items")
|
||||
return f"{title}: {count}"
|
||||
if view_type == "kv":
|
||||
items = view.get("items")
|
||||
count = len(items) if isinstance(items, list) else 0
|
||||
return f"fields: {count}"
|
||||
if view_type == "error":
|
||||
return str(view.get("message") or "error")
|
||||
return ""
|
||||
|
||||
|
||||
def _build_error_payload(code: str, message: str, detail: Any = None) -> Dict[str, Any]:
|
||||
summary = "Canceled" if code == "canceled" else message or "error"
|
||||
view = {"type": "error", "message": summary}
|
||||
payload: Dict[str, Any] = {
|
||||
"ok": False,
|
||||
"summary": summary,
|
||||
"view": view,
|
||||
"error": {
|
||||
"code": code,
|
||||
"message": message,
|
||||
},
|
||||
}
|
||||
if detail is not None:
|
||||
payload["error"]["detail"] = detail
|
||||
return payload
|
||||
|
||||
|
||||
def _normalize_tool_result(result: Any) -> Dict[str, Any]:
|
||||
if isinstance(result, dict) and "ok" in result:
|
||||
payload = dict(result)
|
||||
if payload.get("ok") is False:
|
||||
error = payload.get("error")
|
||||
message = _stringify_value(error.get("message") if isinstance(error, dict) else error)
|
||||
payload.setdefault("summary", message or "error")
|
||||
payload.setdefault("view", {"type": "error", "message": payload["summary"]})
|
||||
return payload
|
||||
data = payload.get("data")
|
||||
if payload.get("view") is None:
|
||||
payload["view"] = _build_view(data)
|
||||
if not payload.get("summary"):
|
||||
payload["summary"] = _build_summary(payload["view"])
|
||||
return payload
|
||||
|
||||
if isinstance(result, dict) and result.get("canceled"):
|
||||
reason = _stringify_value(result.get("reason") or "canceled")
|
||||
return _build_error_payload("canceled", reason, detail=result)
|
||||
|
||||
if isinstance(result, dict) and "error" in result:
|
||||
error = result.get("error")
|
||||
message = _stringify_value(error.get("message") if isinstance(error, dict) else error)
|
||||
return _build_error_payload("error", message, detail=error)
|
||||
|
||||
view = _build_view(result)
|
||||
summary = _build_summary(view)
|
||||
return {"ok": True, "summary": summary, "view": view, "data": result}
|
||||
|
||||
|
||||
def tool_result_to_content(result: Any) -> str:
|
||||
payload = _normalize_tool_result(result)
|
||||
try:
|
||||
return json.dumps(payload, ensure_ascii=False, default=str)
|
||||
except TypeError:
|
||||
return json.dumps({"ok": False, "summary": "error", "view": {"type": "error", "message": "error"}}, ensure_ascii=False)
|
||||
96
domain/agent/tools/processors.py
Normal file
96
domain/agent/tools/processors.py
Normal file
@@ -0,0 +1,96 @@
|
||||
from typing import Any, Dict, Optional
|
||||
|
||||
from domain.processors import ProcessDirectoryRequest, ProcessRequest, ProcessorService
|
||||
from domain.virtual_fs import VirtualFSService
|
||||
|
||||
from .base import ToolSpec
|
||||
|
||||
|
||||
async def _processors_list(_: Dict[str, Any]) -> Dict[str, Any]:
|
||||
return {"processors": ProcessorService.list_processors()}
|
||||
|
||||
|
||||
async def _processors_run(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = str(args.get("path") or "")
|
||||
processor_type = str(args.get("processor_type") or "")
|
||||
config = args.get("config")
|
||||
if not isinstance(config, dict):
|
||||
config = {}
|
||||
|
||||
save_to = args.get("save_to")
|
||||
save_to = str(save_to) if isinstance(save_to, str) and save_to.strip() else None
|
||||
|
||||
max_depth = args.get("max_depth")
|
||||
max_depth_value: Optional[int] = None
|
||||
if max_depth is not None:
|
||||
try:
|
||||
max_depth_value = int(max_depth)
|
||||
except (TypeError, ValueError):
|
||||
max_depth_value = None
|
||||
|
||||
suffix = args.get("suffix")
|
||||
suffix_value = str(suffix) if isinstance(suffix, str) and suffix.strip() else None
|
||||
|
||||
overwrite_value = args.get("overwrite")
|
||||
overwrite = bool(overwrite_value) if overwrite_value is not None else None
|
||||
|
||||
is_dir = await VirtualFSService.path_is_directory(path)
|
||||
if is_dir and (max_depth_value is not None or suffix_value is not None):
|
||||
req = ProcessDirectoryRequest(
|
||||
path=path,
|
||||
processor_type=processor_type,
|
||||
config=config,
|
||||
overwrite=True if overwrite is None else overwrite,
|
||||
max_depth=max_depth_value,
|
||||
suffix=suffix_value,
|
||||
)
|
||||
result = await ProcessorService.process_directory(req)
|
||||
return {"mode": "directory", **result}
|
||||
|
||||
req = ProcessRequest(
|
||||
path=path,
|
||||
processor_type=processor_type,
|
||||
config=config,
|
||||
save_to=save_to,
|
||||
overwrite=False if overwrite is None else overwrite,
|
||||
)
|
||||
result = await ProcessorService.process_file(req)
|
||||
return {"mode": "file", **result}
|
||||
|
||||
|
||||
TOOLS: Dict[str, ToolSpec] = {
|
||||
"processors_list": ToolSpec(
|
||||
name="processors_list",
|
||||
description="获取可用处理器列表（type/name/config_schema 等）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {},
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_processors_list,
|
||||
),
|
||||
"processors_run": ToolSpec(
|
||||
name="processors_run",
|
||||
description=(
|
||||
"运行处理器处理文件或目录。"
|
||||
" 对目录可选 max_depth/suffix；对文件可选 overwrite/save_to。"
|
||||
" 返回任务 id（去任务队列查看进度）。"
|
||||
),
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "文件或目录路径（绝对路径，如 /foo/bar）"},
|
||||
"processor_type": {"type": "string", "description": "处理器类型（例如 image_watermark）"},
|
||||
"config": {"type": "object", "description": "处理器配置，按 processors_list 返回的 config_schema 填写"},
|
||||
"overwrite": {"type": "boolean", "description": "是否覆盖原文件/目录内文件"},
|
||||
"save_to": {"type": "string", "description": "保存到指定路径（仅文件模式，且 overwrite=false 时使用）"},
|
||||
"max_depth": {"type": "integer", "description": "目录遍历深度（仅目录模式）"},
|
||||
"suffix": {"type": "string", "description": "目录批处理时的输出后缀（仅 produces_file 且 overwrite=false）"},
|
||||
},
|
||||
"required": ["path", "processor_type"],
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_processors_run,
|
||||
),
|
||||
}
|
||||
92
domain/agent/tools/time.py
Normal file
92
domain/agent/tools/time.py
Normal file
@@ -0,0 +1,92 @@
|
||||
import calendar
|
||||
from datetime import datetime, timedelta
|
||||
from typing import Any, Dict
|
||||
|
||||
from .base import ToolSpec
|
||||
|
||||
|
||||
def _parse_offset(args: Dict[str, Any], key: str) -> int:
|
||||
value = args.get(key)
|
||||
if value is None:
|
||||
return 0
|
||||
try:
|
||||
return int(value)
|
||||
except (TypeError, ValueError):
|
||||
return 0
|
||||
|
||||
|
||||
def _add_months(dt: datetime, months: int) -> datetime:
|
||||
if months == 0:
|
||||
return dt
|
||||
total = dt.year * 12 + (dt.month - 1) + months
|
||||
year = total // 12
|
||||
month = total % 12 + 1
|
||||
last_day = calendar.monthrange(year, month)[1]
|
||||
day = min(dt.day, last_day)
|
||||
return dt.replace(year=year, month=month, day=day)
|
||||
|
||||
|
||||
async def _time(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
now = datetime.now()
|
||||
year_offset = _parse_offset(args, "year")
|
||||
month_offset = _parse_offset(args, "month")
|
||||
day_offset = _parse_offset(args, "day")
|
||||
hour_offset = _parse_offset(args, "hour")
|
||||
minute_offset = _parse_offset(args, "minute")
|
||||
second_offset = _parse_offset(args, "second")
|
||||
|
||||
dt = _add_months(now, year_offset * 12 + month_offset)
|
||||
dt = dt + timedelta(days=day_offset, hours=hour_offset, minutes=minute_offset, seconds=second_offset)
|
||||
|
||||
weekday_names = [
|
||||
"Monday",
|
||||
"Tuesday",
|
||||
"Wednesday",
|
||||
"Thursday",
|
||||
"Friday",
|
||||
"Saturday",
|
||||
"Sunday",
|
||||
]
|
||||
weekday = weekday_names[dt.weekday()]
|
||||
dt_str = dt.strftime("%Y-%m-%d %H:%M:%S")
|
||||
return {
|
||||
"ok": True,
|
||||
"summary": f"{dt_str} · {weekday}",
|
||||
"data": {
|
||||
"datetime": dt_str,
|
||||
"weekday": weekday,
|
||||
"offset": {
|
||||
"year": year_offset,
|
||||
"month": month_offset,
|
||||
"day": day_offset,
|
||||
"hour": hour_offset,
|
||||
"minute": minute_offset,
|
||||
"second": second_offset,
|
||||
},
|
||||
},
|
||||
}
|
||||
|
||||
|
||||
TOOLS: Dict[str, ToolSpec] = {
|
||||
"time": ToolSpec(
|
||||
name="time",
|
||||
description=(
|
||||
"获取服务器当前时间（精确到秒，含英文星期）。"
|
||||
" 支持 year/month/day/hour/minute/second 偏移（可为负数）。"
|
||||
),
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"year": {"type": "integer", "description": "年偏移（可为负数）"},
|
||||
"month": {"type": "integer", "description": "月偏移（可为负数）"},
|
||||
"day": {"type": "integer", "description": "日偏移（可为负数）"},
|
||||
"hour": {"type": "integer", "description": "时偏移（可为负数）"},
|
||||
"minute": {"type": "integer", "description": "分偏移（可为负数）"},
|
||||
"second": {"type": "integer", "description": "秒偏移（可为负数）"},
|
||||
},
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_time,
|
||||
),
|
||||
}
|
||||
287
domain/agent/tools/vfs.py
Normal file
287
domain/agent/tools/vfs.py
Normal file
@@ -0,0 +1,287 @@
|
||||
from typing import Any, Dict, Optional
|
||||
|
||||
from domain.virtual_fs import VirtualFSService
|
||||
from domain.virtual_fs.search import VirtualFSSearchService
|
||||
|
||||
from .base import ToolSpec
|
||||
|
||||
|
||||
def _normalize_vfs_path(value: Any) -> str:
|
||||
s = str(value or "").strip().replace("\\", "/")
|
||||
if not s:
|
||||
return ""
|
||||
if not s.startswith("/"):
|
||||
s = "/" + s
|
||||
s = s.rstrip("/") or "/"
|
||||
return s
|
||||
|
||||
|
||||
def _require_vfs_path(value: Any, field: str) -> str:
|
||||
path = _normalize_vfs_path(value)
|
||||
if not path:
|
||||
raise ValueError(f"missing_{field}")
|
||||
return path
|
||||
|
||||
|
||||
async def _vfs_list_dir(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _normalize_vfs_path(args.get("path") or "/") or "/"
|
||||
page = int(args.get("page") or 1)
|
||||
page_size = int(args.get("page_size") or 50)
|
||||
sort_by = str(args.get("sort_by") or "name")
|
||||
sort_order = str(args.get("sort_order") or "asc")
|
||||
return await VirtualFSService.list_directory(path, page, page_size, sort_by, sort_order)
|
||||
|
||||
|
||||
async def _vfs_stat(args: Dict[str, Any]) -> Any:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
return await VirtualFSService.stat(path)
|
||||
|
||||
|
||||
async def _vfs_read_text(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
encoding = str(args.get("encoding") or "utf-8")
|
||||
max_chars = int(args.get("max_chars") or 8000)
|
||||
|
||||
data = await VirtualFSService.read_file(path)
|
||||
if isinstance(data, (bytes, bytearray)):
|
||||
try:
|
||||
text = bytes(data).decode(encoding)
|
||||
except UnicodeDecodeError:
|
||||
return {"error": "binary_or_invalid_text", "path": path}
|
||||
elif isinstance(data, str):
|
||||
text = data
|
||||
else:
|
||||
text = str(data)
|
||||
|
||||
original_len = len(text)
|
||||
truncated = original_len > max_chars
|
||||
if truncated:
|
||||
text = text[:max_chars]
|
||||
return {
|
||||
"path": path,
|
||||
"encoding": encoding,
|
||||
"content": text,
|
||||
"truncated": truncated,
|
||||
"length": original_len,
|
||||
}
|
||||
|
||||
|
||||
async def _vfs_write_text(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
if path == "/":
|
||||
raise ValueError("invalid_path")
|
||||
encoding = str(args.get("encoding") or "utf-8")
|
||||
content = str(args.get("content") or "")
|
||||
data = content.encode(encoding)
|
||||
await VirtualFSService.write_file(path, data)
|
||||
return {"written": True, "path": path, "encoding": encoding, "bytes": len(data)}
|
||||
|
||||
|
||||
async def _vfs_mkdir(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
return await VirtualFSService.mkdir(path)
|
||||
|
||||
|
||||
async def _vfs_delete(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
path = _require_vfs_path(args.get("path"), "path")
|
||||
return await VirtualFSService.delete(path)
|
||||
|
||||
|
||||
async def _vfs_move(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
src = _require_vfs_path(args.get("src"), "src")
|
||||
dst = _require_vfs_path(args.get("dst"), "dst")
|
||||
if src == "/" or dst == "/":
|
||||
raise ValueError("invalid_path")
|
||||
overwrite = bool(args.get("overwrite") or False)
|
||||
return await VirtualFSService.move(src, dst, overwrite)
|
||||
|
||||
|
||||
async def _vfs_copy(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
src = _require_vfs_path(args.get("src"), "src")
|
||||
dst = _require_vfs_path(args.get("dst"), "dst")
|
||||
if src == "/" or dst == "/":
|
||||
raise ValueError("invalid_path")
|
||||
overwrite = bool(args.get("overwrite") or False)
|
||||
return await VirtualFSService.copy(src, dst, overwrite)
|
||||
|
||||
|
||||
async def _vfs_rename(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
src = _require_vfs_path(args.get("src"), "src")
|
||||
dst = _require_vfs_path(args.get("dst"), "dst")
|
||||
if src == "/" or dst == "/":
|
||||
raise ValueError("invalid_path")
|
||||
overwrite = bool(args.get("overwrite") or False)
|
||||
return await VirtualFSService.rename(src, dst, overwrite)
|
||||
|
||||
|
||||
async def _vfs_search(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
q = str(args.get("q") or "").strip()
|
||||
if not q:
|
||||
raise ValueError("missing_q")
|
||||
mode = str(args.get("mode") or "vector")
|
||||
top_k = int(args.get("top_k") or 10)
|
||||
page = int(args.get("page") or 1)
|
||||
page_size = int(args.get("page_size") or 10)
|
||||
return await VirtualFSSearchService.search(q, top_k, mode, page, page_size)
|
||||
|
||||
|
||||
TOOLS: Dict[str, ToolSpec] = {
|
||||
"vfs_list_dir": ToolSpec(
|
||||
name="vfs_list_dir",
|
||||
description="浏览目录（列出 entries + pagination）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "目录路径（绝对路径，如 /foo/bar）"},
|
||||
"page": {"type": "integer", "description": "页码（从 1 开始）"},
|
||||
"page_size": {"type": "integer", "description": "每页条数"},
|
||||
"sort_by": {"type": "string", "description": "排序字段：name/size/mtime"},
|
||||
"sort_order": {"type": "string", "description": "排序顺序：asc/desc"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_list_dir,
|
||||
),
|
||||
"vfs_stat": ToolSpec(
|
||||
name="vfs_stat",
|
||||
description="查看文件/目录信息（size/mtime/is_dir/has_thumbnail/vector_index 等）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "路径（绝对路径，如 /foo/bar.txt）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_stat,
|
||||
),
|
||||
"vfs_read_text": ToolSpec(
|
||||
name="vfs_read_text",
|
||||
description="读取文本文件内容（解码失败视为二进制，返回 error）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "文件路径（绝对路径，如 /foo/bar.md）"},
|
||||
"encoding": {"type": "string", "description": "文本编码（默认 utf-8）"},
|
||||
"max_chars": {"type": "integer", "description": "最多返回的字符数（默认 8000）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_read_text,
|
||||
),
|
||||
"vfs_write_text": ToolSpec(
|
||||
name="vfs_write_text",
|
||||
description="写入文本文件内容（会覆盖目标文件）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "文件路径（绝对路径，如 /foo/bar.md）"},
|
||||
"content": {"type": "string", "description": "要写入的文本内容"},
|
||||
"encoding": {"type": "string", "description": "文本编码（默认 utf-8）"},
|
||||
},
|
||||
"required": ["path", "content"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_write_text,
|
||||
),
|
||||
"vfs_mkdir": ToolSpec(
|
||||
name="vfs_mkdir",
|
||||
description="创建目录。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "目录路径（绝对路径，如 /foo/bar）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_mkdir,
|
||||
),
|
||||
"vfs_delete": ToolSpec(
|
||||
name="vfs_delete",
|
||||
description="删除文件或目录（由底层适配器决定是否递归）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"path": {"type": "string", "description": "路径（绝对路径，如 /foo/bar 或 /foo/bar.txt）"},
|
||||
},
|
||||
"required": ["path"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_delete,
|
||||
),
|
||||
"vfs_move": ToolSpec(
|
||||
name="vfs_move",
|
||||
description="移动路径（可能进入任务队列）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"src": {"type": "string", "description": "源路径（绝对路径）"},
|
||||
"dst": {"type": "string", "description": "目标路径（绝对路径）"},
|
||||
"overwrite": {"type": "boolean", "description": "是否允许覆盖已存在目标（默认 false）"},
|
||||
},
|
||||
"required": ["src", "dst"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_move,
|
||||
),
|
||||
"vfs_copy": ToolSpec(
|
||||
name="vfs_copy",
|
||||
description="复制路径（可能进入任务队列）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"src": {"type": "string", "description": "源路径（绝对路径）"},
|
||||
"dst": {"type": "string", "description": "目标路径（绝对路径）"},
|
||||
"overwrite": {"type": "boolean", "description": "是否覆盖已存在目标（默认 false）"},
|
||||
},
|
||||
"required": ["src", "dst"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_copy,
|
||||
),
|
||||
"vfs_rename": ToolSpec(
|
||||
name="vfs_rename",
|
||||
description="重命名路径（本质是同目录 move）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"src": {"type": "string", "description": "源路径（绝对路径）"},
|
||||
"dst": {"type": "string", "description": "目标路径（绝对路径）"},
|
||||
"overwrite": {"type": "boolean", "description": "是否允许覆盖已存在目标（默认 false）"},
|
||||
},
|
||||
"required": ["src", "dst"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=True,
|
||||
handler=_vfs_rename,
|
||||
),
|
||||
"vfs_search": ToolSpec(
|
||||
name="vfs_search",
|
||||
description="搜索文件（mode=vector 或 filename）。",
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"q": {"type": "string", "description": "搜索关键词"},
|
||||
"mode": {"type": "string", "description": "搜索模式：vector/filename（默认 vector）"},
|
||||
"top_k": {"type": "integer", "description": "返回数量（vector 模式使用，默认 10）"},
|
||||
"page": {"type": "integer", "description": "页码（filename 模式使用，默认 1）"},
|
||||
"page_size": {"type": "integer", "description": "分页大小（filename 模式使用，默认 10）"},
|
||||
},
|
||||
"required": ["q"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_vfs_search,
|
||||
),
|
||||
}
|
||||
182
domain/agent/tools/web_fetch.py
Normal file
182
domain/agent/tools/web_fetch.py
Normal file
@@ -0,0 +1,182 @@
|
||||
from html.parser import HTMLParser
|
||||
from typing import Any, Dict, List
|
||||
from urllib.parse import urljoin
|
||||
|
||||
import httpx
|
||||
|
||||
from .base import ToolSpec
|
||||
|
||||
|
||||
class _HtmlTextExtractor(HTMLParser):
|
||||
def __init__(self, base_url: str):
|
||||
super().__init__()
|
||||
self.base_url = base_url
|
||||
self.links: List[str] = []
|
||||
self._link_set: set[str] = set()
|
||||
self._title_parts: List[str] = []
|
||||
self._text_parts: List[str] = []
|
||||
self._in_title = False
|
||||
self._skip_text = False
|
||||
|
||||
def handle_starttag(self, tag: str, attrs: List[tuple[str, str | None]]):
|
||||
tag = tag.lower()
|
||||
if tag == "title":
|
||||
self._in_title = True
|
||||
if tag in ("script", "style", "noscript"):
|
||||
self._skip_text = True
|
||||
if tag != "a":
|
||||
return
|
||||
href = ""
|
||||
for key, value in attrs:
|
||||
if key.lower() == "href":
|
||||
href = str(value or "").strip()
|
||||
break
|
||||
if not href or href.startswith("#"):
|
||||
return
|
||||
lower = href.lower()
|
||||
if lower.startswith(("javascript:", "mailto:", "tel:", "data:")):
|
||||
return
|
||||
resolved = urljoin(self.base_url, href)
|
||||
if resolved in self._link_set:
|
||||
return
|
||||
self._link_set.add(resolved)
|
||||
self.links.append(resolved)
|
||||
|
||||
def handle_endtag(self, tag: str):
|
||||
tag = tag.lower()
|
||||
if tag == "title":
|
||||
self._in_title = False
|
||||
if tag in ("script", "style", "noscript"):
|
||||
self._skip_text = False
|
||||
|
||||
def handle_data(self, data: str):
|
||||
if not data:
|
||||
return
|
||||
if self._in_title:
|
||||
self._title_parts.append(data)
|
||||
if self._skip_text:
|
||||
return
|
||||
if data.strip():
|
||||
self._text_parts.append(data)
|
||||
|
||||
@property
|
||||
def title(self) -> str:
|
||||
return " ".join(part.strip() for part in self._title_parts if part and part.strip()).strip()
|
||||
|
||||
@property
|
||||
def text(self) -> str:
|
||||
if not self._text_parts:
|
||||
return ""
|
||||
text = " ".join(part.strip() for part in self._text_parts if part and part.strip())
|
||||
return " ".join(text.split())
|
||||
|
||||
|
||||
async def _web_fetch(args: Dict[str, Any]) -> Dict[str, Any]:
|
||||
url = str(args.get("url") or "").strip()
|
||||
if not url:
|
||||
raise ValueError("missing_url")
|
||||
|
||||
method = str(args.get("method") or "GET").upper()
|
||||
allowed_methods = {"GET", "POST", "PUT", "PATCH", "DELETE", "HEAD", "OPTIONS"}
|
||||
if method not in allowed_methods:
|
||||
raise ValueError("invalid_method")
|
||||
|
||||
headers_raw = args.get("headers")
|
||||
headers = {str(k): str(v) for k, v in headers_raw.items() if v is not None} if isinstance(headers_raw, dict) else None
|
||||
params_raw = args.get("params")
|
||||
params = {str(k): str(v) for k, v in params_raw.items() if v is not None} if isinstance(params_raw, dict) else None
|
||||
json_body = args.get("json") if "json" in args else None
|
||||
body = args.get("body")
|
||||
|
||||
request_kwargs: Dict[str, Any] = {}
|
||||
if headers:
|
||||
request_kwargs["headers"] = headers
|
||||
if params:
|
||||
request_kwargs["params"] = params
|
||||
if json_body is not None:
|
||||
request_kwargs["json"] = json_body
|
||||
elif body is not None:
|
||||
request_kwargs["content"] = str(body)
|
||||
|
||||
async with httpx.AsyncClient(timeout=20.0, follow_redirects=True) as client:
|
||||
resp = await client.request(method, url, **request_kwargs)
|
||||
|
||||
content_type = resp.headers.get("content-type") or ""
|
||||
text = resp.text or ""
|
||||
is_html = "html" in content_type.lower()
|
||||
if not is_html:
|
||||
probe = text.lstrip()[:200].lower()
|
||||
if "<html" in probe or "<!doctype html" in probe:
|
||||
is_html = True
|
||||
|
||||
html = ""
|
||||
title = ""
|
||||
links: List[str] = []
|
||||
extracted_text = text
|
||||
|
||||
if is_html and text:
|
||||
html = text
|
||||
parser = _HtmlTextExtractor(str(resp.url))
|
||||
parser.feed(text)
|
||||
title = parser.title
|
||||
links = parser.links
|
||||
extracted_text = parser.text
|
||||
|
||||
data = {
|
||||
"url": url,
|
||||
"method": method,
|
||||
"final_url": str(resp.url),
|
||||
"status_code": resp.status_code,
|
||||
"content_type": content_type,
|
||||
"title": title,
|
||||
"html": html,
|
||||
"text": extracted_text,
|
||||
"links": links,
|
||||
}
|
||||
|
||||
summary_parts = [method, str(resp.status_code)]
|
||||
if title:
|
||||
summary_parts.append(title)
|
||||
summary_parts.append(f"{len(links)} links")
|
||||
summary = " · ".join(summary_parts)
|
||||
|
||||
view = {
|
||||
"type": "text",
|
||||
"text": extracted_text,
|
||||
"meta": {
|
||||
"url": url,
|
||||
"final_url": str(resp.url),
|
||||
"status_code": resp.status_code,
|
||||
"content_type": content_type,
|
||||
"title": title,
|
||||
"method": method,
|
||||
"links": len(links),
|
||||
},
|
||||
}
|
||||
return {"ok": True, "summary": summary, "view": view, "data": data}
|
||||
|
||||
|
||||
TOOLS: Dict[str, ToolSpec] = {
|
||||
"web_fetch": ToolSpec(
|
||||
name="web_fetch",
|
||||
description=(
|
||||
"抓取网页内容，返回状态、标题、正文、HTML、链接等信息。"
|
||||
" 支持 GET/POST/PUT/PATCH/DELETE/HEAD/OPTIONS。"
|
||||
),
|
||||
parameters={
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"url": {"type": "string", "description": "目标 URL"},
|
||||
"method": {"type": "string", "description": "请求方法（默认 GET）"},
|
||||
"headers": {"type": "object", "description": "请求头", "additionalProperties": {"type": "string"}},
|
||||
"params": {"type": "object", "description": "查询参数", "additionalProperties": {"type": "string"}},
|
||||
"json": {"type": "object", "description": "JSON 请求体"},
|
||||
"body": {"type": "string", "description": "原始请求体"},
|
||||
},
|
||||
"required": ["url"],
|
||||
"additionalProperties": False,
|
||||
},
|
||||
requires_confirmation=False,
|
||||
handler=_web_fetch,
|
||||
),
|
||||
}
|
||||
Reference in New Issue
Block a user