Merge branch 'master' of http://192.168.0.200/knowledge-base-bj/deerflow-server into xhy
This commit is contained in:
commit
dadb0bbe68
782 files changed
+160067
-3424
No files matched your search
+71
@@ -0,0 +1,71 @@
|
||||
"""笔记本 RAG 检索工具。
|
||||
|
||||
从 LangGraph configurable 中读取 notebook_search_space_id / notebook_source_ids /
|
||||
notebook_token,调用外部检索接口返回相关文本块,供 LLM 在笔记本模式下引用回答。
|
||||
|
||||
API base URL 通过环境变量 DEER_FLOW_NOTEBOOK_API_BASE 配置。
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import logging
|
||||
import os
|
||||
|
||||
import httpx
|
||||
from langchain.tools import tool
|
||||
from langgraph.config import get_config
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
_SEARCH_PATH = "/api/v1/source/mentioned-documents/search"
|
||||
|
||||
|
||||
def _notebook_api_base() -> str:
|
||||
return (os.getenv("DEER_FLOW_NOTEBOOK_API_BASE") or "").rstrip("/")
|
||||
|
||||
|
||||
@tool("search_notebook_sources", parse_docstring=True)
|
||||
def notebook_search_tool(query: str, top_k: int = 10) -> str:
|
||||
"""Search the notebook's knowledge sources to find relevant content.
|
||||
|
||||
Call this tool for EVERY user question. After retrieving results, cite
|
||||
sources inline using [[cite:<document_id>]] markers immediately after
|
||||
each claim, e.g. "违约金为0.5%[[cite:101]]".
|
||||
|
||||
Args:
|
||||
query: The user's question or a refined search query.
|
||||
top_k: Number of top chunks to retrieve (default 10).
|
||||
"""
|
||||
base = _notebook_api_base()
|
||||
if not base:
|
||||
return "Notebook API base URL not configured (DEER_FLOW_NOTEBOOK_API_BASE)."
|
||||
|
||||
cfg = get_config().get("configurable", {})
|
||||
search_space_id: int | None = cfg.get("notebook_search_space_id")
|
||||
source_ids: list[int] = cfg.get("notebook_source_ids") or []
|
||||
token: str = cfg.get("notebook_token") or ""
|
||||
|
||||
if not search_space_id:
|
||||
return "No notebook search space configured."
|
||||
|
||||
logger.info("notebook_search_tool query=%r source_ids=%s search_space_id=%s", query, source_ids, search_space_id)
|
||||
|
||||
try:
|
||||
resp = httpx.post(
|
||||
f"{base}{_SEARCH_PATH}",
|
||||
headers={"Authorization": f"Bearer {token}"},
|
||||
json={
|
||||
"query": query,
|
||||
"source_ids": [source_ids],
|
||||
"search_space_id": search_space_id,
|
||||
"top_k": top_k,
|
||||
},
|
||||
timeout=30,
|
||||
)
|
||||
resp.raise_for_status()
|
||||
data = resp.json()
|
||||
content = data.get("content") if isinstance(data, dict) else None
|
||||
return content if content else "No relevant content found."
|
||||
except Exception as e:
|
||||
logger.warning("notebook_search_tool failed: %s", e)
|
||||
return f"Search failed: {e}"
|
||||
Reference in new issue
Block a user