mirror of
https://github.com/ThearchyHelios/AgentLab.git
synced 2026-11-15 04:21:57 +00:00
ThearchyHelios
86a0928512
本机配的是 LM Studio 上的 text-embedding-qwen3-embedding-4b。服务启动时它没起， load_setting 连不上，整个进程就一直用本地哈希——它后来起了也不会自己连上，原因只在 服务端日志里留一句。页面按"配的是不是 local"判断，于是： - 不标"没有语义能力"，看着一切正常，检索却一直是纯关键词； - 高亮"重建索引（73 段）"——那些"对不上"的向量正是 qwen3 建好的，一点就拿哈希 覆盖掉，连上之后还得再重建一遍。 改法： 1. 后端记下保存的配置为什么没生效（embeddings.unavailable_reason），状态接口多报 fallback / fallback_reason：配的是语义模型、实际在用本地哈希。重新配置成功就清掉。 2. 页面按实际生效的 embedder（has_semantics）标"没有语义能力"。退回时说清配的是哪个、 在哪、为什么没连上，给「重新连接」——发的就是保存着的那份配置，服务起来后点它即可， 不用重启；"重建索引"收起，存量提示改口成先别重建。明确选了本地哈希的照旧：标没有 语义能力，重建按钮还在。 验证：test_kb 新增一条，走完"启动连不上 → 状态里报退回并带原因 → 重新连上后消失 → 明确选本地不算退回"。check-ui 新增一组伪造状态接口的断言（退回 / 重新连接 / 明确选 本地，写请求一律拦下）；把页面换回旧版本跑，这组挂 8 项。后端 471 passed，tsc 与 check-decode / guards / stream / ui 全绿；本机状态接口现在如实报出退回和原因。
461 lines
17 KiB
Python
461 lines
17 KiB
Python
from __future__ import annotations
|
||
|
||
from typing import Any
|
||
|
||
from fastapi import (
|
||
APIRouter, BackgroundTasks, Depends, File, HTTPException, Query, UploadFile,
|
||
)
|
||
from pydantic import BaseModel, Field
|
||
from sqlalchemy import select
|
||
from sqlalchemy.ext.asyncio import AsyncSession
|
||
|
||
from app.core.config import settings
|
||
from app.db.base import get_session
|
||
from app.db.models import Document, MemoryItem, Skill
|
||
from app.memory import inverted, kb, store
|
||
|
||
# --------------------------------------------------------------------------
|
||
# 长期记忆
|
||
# --------------------------------------------------------------------------
|
||
|
||
memory_router = APIRouter(prefix="/api/memory", tags=["memory"])
|
||
|
||
|
||
class MemoryIn(BaseModel):
|
||
content: str = Field(min_length=1)
|
||
scope: str = "default"
|
||
kind: str = "fact"
|
||
importance: float = Field(default=0.5, ge=0, le=1)
|
||
meta: dict[str, Any] = Field(default_factory=dict)
|
||
|
||
|
||
class MemoryOut(BaseModel):
|
||
id: str
|
||
scope: str
|
||
kind: str
|
||
content: str
|
||
importance: float
|
||
use_count: int
|
||
meta: dict[str, Any]
|
||
created_at: Any = None
|
||
|
||
model_config = {"from_attributes": True}
|
||
|
||
|
||
@memory_router.get("/scopes")
|
||
async def memory_scopes(session: AsyncSession = Depends(get_session)) -> list[dict[str, Any]]:
|
||
return await store.list_scopes(session)
|
||
|
||
|
||
@memory_router.get("", response_model=list[MemoryOut])
|
||
async def list_memory(
|
||
scope: str | None = None,
|
||
limit: int = Query(default=200, le=500),
|
||
session: AsyncSession = Depends(get_session),
|
||
) -> list[MemoryItem]:
|
||
return await store.list_memories(session, scope=scope, limit=limit)
|
||
|
||
|
||
@memory_router.post("", response_model=MemoryOut, status_code=201)
|
||
async def add_memory(
|
||
payload: MemoryIn, session: AsyncSession = Depends(get_session)
|
||
) -> MemoryItem:
|
||
return await store.remember(
|
||
session,
|
||
scope=payload.scope,
|
||
content=payload.content,
|
||
kind=payload.kind,
|
||
importance=payload.importance,
|
||
meta=payload.meta,
|
||
)
|
||
|
||
|
||
@memory_router.get("/search")
|
||
async def search_memory(
|
||
q: str,
|
||
scope: str = "default",
|
||
limit: int = 10,
|
||
session: AsyncSession = Depends(get_session),
|
||
) -> dict[str, Any]:
|
||
hits = await store.recall(session, scope=scope, query=q, limit=limit)
|
||
return {"query": q, "scope": scope, "results": hits}
|
||
|
||
|
||
@memory_router.delete("/{memory_id}", status_code=204)
|
||
async def delete_memory(
|
||
memory_id: str, session: AsyncSession = Depends(get_session)
|
||
) -> None:
|
||
if not await store.forget(session, memory_id):
|
||
raise HTTPException(404, "记忆不存在")
|
||
|
||
|
||
@memory_router.delete("/scope/{scope}")
|
||
async def clear_scope(scope: str, session: AsyncSession = Depends(get_session)) -> dict[str, int]:
|
||
return {"removed": await store.clear_scope(session, scope)}
|
||
|
||
|
||
# --------------------------------------------------------------------------
|
||
# 知识库
|
||
# --------------------------------------------------------------------------
|
||
|
||
kb_router = APIRouter(prefix="/api/kb", tags=["knowledge"])
|
||
|
||
|
||
class DocumentOut(BaseModel):
|
||
id: str
|
||
collection: str
|
||
title: str
|
||
source: str
|
||
mime: str
|
||
chunk_count: int
|
||
#: ready | processing | failed。前端据此显示"处理中 3200/7914 段"
|
||
status: str = "ready"
|
||
error: str = ""
|
||
meta: dict[str, Any] = Field(default_factory=dict)
|
||
created_at: Any = None
|
||
|
||
model_config = {"from_attributes": True}
|
||
|
||
|
||
@kb_router.get("/collections")
|
||
async def collections(session: AsyncSession = Depends(get_session)) -> list[dict[str, Any]]:
|
||
return await kb.list_collections(session)
|
||
|
||
|
||
@kb_router.get("/documents", response_model=list[DocumentOut])
|
||
async def list_documents(
|
||
collection: str | None = None,
|
||
session: AsyncSession = Depends(get_session),
|
||
) -> list[Document]:
|
||
stmt = select(Document).order_by(Document.created_at.desc()).limit(300)
|
||
if collection:
|
||
stmt = stmt.where(Document.collection == collection)
|
||
return list((await session.execute(stmt)).scalars())
|
||
|
||
|
||
class IngestIn(BaseModel):
|
||
collection: str = "default"
|
||
title: str = ""
|
||
content: str = Field(min_length=1)
|
||
source: str = ""
|
||
|
||
|
||
@kb_router.post("/documents", response_model=DocumentOut, status_code=201)
|
||
async def ingest(payload: IngestIn, session: AsyncSession = Depends(get_session)) -> Document:
|
||
return await kb.ingest_document(
|
||
session,
|
||
collection=payload.collection,
|
||
title=payload.title,
|
||
content=payload.content,
|
||
source=payload.source,
|
||
)
|
||
|
||
|
||
@kb_router.post("/upload", response_model=DocumentOut, status_code=201)
|
||
async def upload(
|
||
background: BackgroundTasks,
|
||
file: UploadFile = File(...),
|
||
collection: str = "default",
|
||
session: AsyncSession = Depends(get_session),
|
||
) -> Document:
|
||
# 边读边数，超了立刻停——原来是 raw = await file.read() 读完再判断，
|
||
# 传个 1GB 的文件，内存在报 413 之前就已经吃掉了。那个检查拦的是"入库"，
|
||
# 不是"占内存"
|
||
limit = settings.max_upload_mb * 1024 * 1024
|
||
chunks: list[bytes] = []
|
||
total = 0
|
||
while piece := await file.read(1024 * 1024):
|
||
total += len(piece)
|
||
if total > limit:
|
||
raise HTTPException(413, f"文件超过 {settings.max_upload_mb}MB")
|
||
chunks.append(piece)
|
||
raw = b"".join(chunks)
|
||
|
||
from app.memory.parsing import UnsupportedDocument, extract
|
||
|
||
try:
|
||
content = extract(raw, file.filename or "", file.content_type or "")
|
||
except UnsupportedDocument as e:
|
||
# 原样交回：里面写的是"装哪个包"或"这是扫描件"，都能照着做
|
||
raise HTTPException(415, str(e)) from None
|
||
doc = await kb.create_document(
|
||
session,
|
||
collection=collection,
|
||
title=file.filename or "上传文档",
|
||
content=content,
|
||
source=file.filename or "",
|
||
mime=file.content_type or "text/plain",
|
||
status="processing",
|
||
)
|
||
# 立刻返回，切块和算向量在后台跑。实测一个 10MB 文档用本地哈希要两分半，
|
||
# 换成远端 embedding 是几十分钟——压在请求里必然超时，而后端其实还在跑，
|
||
# 界面显示失败、数据其实成功，比直接拒绝还糟
|
||
background.add_task(_process_in_background, doc.id)
|
||
return doc
|
||
|
||
|
||
async def _process_in_background(doc_id: str) -> None:
|
||
"""后台切块。进度写回文档行，失败也写回——不然它会永远停在"处理中"。"""
|
||
from app.db.base import SessionLocal
|
||
|
||
async with SessionLocal() as session:
|
||
doc = await session.get(Document, doc_id)
|
||
if not doc:
|
||
return
|
||
try:
|
||
# 进度由 process_document 自己按批提交——它就在主事务里，
|
||
# 不会和另一个 session 抢 SQLite 的写锁
|
||
await kb.process_document(session, doc)
|
||
except Exception as e: # noqa: BLE001
|
||
await session.rollback()
|
||
doc = await session.get(Document, doc_id)
|
||
if doc:
|
||
doc.status = "failed"
|
||
doc.error = f"{type(e).__name__}: {e}"[:500]
|
||
await session.commit()
|
||
|
||
|
||
@kb_router.get("/search")
|
||
async def search_kb(
|
||
q: str,
|
||
collection: str | None = None,
|
||
limit: int = 5,
|
||
alpha: float | None = Query(default=None, ge=0, le=1,
|
||
description="1=纯向量, 0=纯关键词；不传则按 embedder 能力取默认"),
|
||
session: AsyncSession = Depends(get_session),
|
||
) -> dict[str, Any]:
|
||
notes: list[str] = []
|
||
hits = await kb.search(session, collection=collection, query=q, limit=limit,
|
||
alpha=alpha, on_degrade=notes.append)
|
||
# 退回关键词这件事要跟着结果一起交出去，而不是只写进服务端日志——
|
||
# 调用方（设置页的"试一下"、外部脚本）看到的是结果变差，得知道为什么
|
||
return {"query": q, "collection": collection, "results": hits, "degraded": notes}
|
||
|
||
|
||
@kb_router.get("/embedding")
|
||
async def embedding_status(
|
||
collection: str | None = None, session: AsyncSession = Depends(get_session)
|
||
) -> dict[str, Any]:
|
||
"""当前用的是哪个 embedder，以及有多少条向量已经对不上了。"""
|
||
from app.memory.embeddings import (
|
||
EMBEDDING_SETTING_KEY, default_alpha, embedder_dim, embedder_id, has_semantics,
|
||
unavailable_reason,
|
||
)
|
||
from app.db.models import Setting
|
||
|
||
row = await session.get(Setting, EMBEDDING_SETTING_KEY)
|
||
saved = (row.value if row else {}) or {}
|
||
# 配的是语义模型，实际在用本地哈希：保存的配置没生效（多半是启动时本机的模型
|
||
# 服务还没起）。这时"对不上"的存量向量正是那个模型建好的，界面不能再劝人重建——
|
||
# 那等于拿哈希把它们覆盖掉，连上之后还得再重建一遍
|
||
fallback = saved.get("kind") == "openai" and not has_semantics()
|
||
return {
|
||
"embedder": embedder_id(),
|
||
"dim": embedder_dim(),
|
||
"configured": bool(saved.get("kind")),
|
||
"kind": saved.get("kind", "local"),
|
||
"model": saved.get("model", ""),
|
||
"base_url": saved.get("base_url", ""),
|
||
"stale_chunks": await kb.stale_count(session, collection),
|
||
# 记忆和知识库共用一个 embedder，换模型时一起失效——报一个漏一个
|
||
# 的话，用户点完重建还是想不起事，而且不知道为什么
|
||
"stale_memories": await store.stale_count(session),
|
||
"has_semantics": has_semantics(),
|
||
"fallback": fallback,
|
||
"fallback_reason": (unavailable_reason() or "") if fallback else "",
|
||
"default_alpha": default_alpha(),
|
||
# 没建倒排的片段会走全表扫——结果对，但慢
|
||
"unindexed_chunks": await inverted.missing_count(session, collection),
|
||
}
|
||
|
||
|
||
class EmbeddingIn(BaseModel):
|
||
kind: str = Field(default="local", pattern="^(local|openai)$")
|
||
model: str = ""
|
||
#: 任何讲 OpenAI /v1/embeddings 的地址：本机 LM Studio / Ollama / vLLM / TEI，
|
||
#: 或者你自己的网关。留空才走 api.openai.com
|
||
base_url: str = ""
|
||
|
||
|
||
@kb_router.put("/embedding")
|
||
async def set_embedding(
|
||
payload: EmbeddingIn, session: AsyncSession = Depends(get_session)
|
||
) -> dict[str, Any]:
|
||
"""换 embedding 模型。
|
||
|
||
建不起来就 400 原样交回原因——选了远端却静默退回本地哈希向量，
|
||
用户会以为自己在用语义检索，而搜不出同义表达时完全无从怀疑到这里。
|
||
"""
|
||
from app.memory import embeddings as emb
|
||
from app.db.models import Setting
|
||
|
||
try:
|
||
emb.configure(payload.kind, payload.model, payload.base_url)
|
||
except emb.EmbedderUnavailable as e:
|
||
raise HTTPException(400, str(e)) from e
|
||
|
||
row = await session.get(Setting, emb.EMBEDDING_SETTING_KEY)
|
||
value = {"kind": payload.kind, "model": payload.model,
|
||
"base_url": payload.base_url}
|
||
if row:
|
||
row.value = value
|
||
else:
|
||
session.add(Setting(key=emb.EMBEDDING_SETTING_KEY, value=value))
|
||
await session.commit()
|
||
return {
|
||
"embedder": emb.embedder_id(),
|
||
"dim": emb.embedder_dim(),
|
||
"stale_chunks": await kb.stale_count(session),
|
||
"stale_memories": await store.stale_count(session),
|
||
}
|
||
|
||
|
||
@kb_router.get("/embedding/probe")
|
||
async def probe_embedding(base_url: str) -> dict[str, Any]:
|
||
"""问问这个地址上有哪些模型。
|
||
|
||
自己填模型名太容易写错（LM Studio 里叫
|
||
text-embedding-qwen3-embedding-4b，不是 Qwen3-Embedding-4B），
|
||
而填错的表现是切换时报一句 404，人还以为是服务没起。
|
||
"""
|
||
import httpx
|
||
|
||
url = base_url.rstrip("/") + "/models"
|
||
try:
|
||
async with httpx.AsyncClient(timeout=8) as client:
|
||
data = (await client.get(url)).json()
|
||
except Exception as e: # noqa: BLE001
|
||
raise HTTPException(400, f"连不上 {url}：{type(e).__name__}: {e}") from e
|
||
models = [m.get("id", "") for m in (data.get("data") or []) if m.get("id")]
|
||
return {"base_url": base_url, "models": models}
|
||
|
||
|
||
@kb_router.post("/reindex")
|
||
async def reindex(
|
||
collection: str | None = None, session: AsyncSession = Depends(get_session)
|
||
) -> dict[str, Any]:
|
||
"""用当前 embedder 重算向量。换了模型之后唯一的恢复手段。
|
||
|
||
知识库和长期记忆一起重建：它们共用一个 embedder，换模型时一起失效。
|
||
分成两个按钮的结果是点了一个、以为好了，另一个还在悄悄退回关键词——
|
||
真踩过，记忆那边连个提示都没有。
|
||
"""
|
||
out = await kb.reindex(session, collection)
|
||
# 记忆不按 collection 分，重建就是全量
|
||
out["memories_reindexed"] = await store.reindex(session)
|
||
return out
|
||
|
||
|
||
#: 单段返回多少字。切块本来就在千字量级，详情页是给人扫一眼"切得对不对"的，
|
||
#: 不是给人通读原文的——整篇几千段全量吐出去，一份大文档能有几兆
|
||
_CHUNK_PREVIEW = 1200
|
||
|
||
|
||
@kb_router.get("/documents/{doc_id}")
|
||
async def get_document(
|
||
doc_id: str, session: AsyncSession = Depends(get_session)
|
||
) -> dict[str, Any]:
|
||
"""一份文档被切成了什么样。
|
||
|
||
切块这一层做了不少事——按行边界留重叠、跨块把表头续上（memory/kb.py 的
|
||
chunk_text），而在此之前结果在界面上一处都看不见。检索不准的时候第一个
|
||
该看的就是它：表头有没有续上、重叠对不对、哪一段被截断了。
|
||
"""
|
||
from app.db.models import Chunk
|
||
|
||
doc = await session.get(Document, doc_id)
|
||
if not doc:
|
||
raise HTTPException(404, "文档不存在")
|
||
|
||
rows = list((await session.execute(
|
||
select(Chunk).where(Chunk.document_id == doc_id).order_by(Chunk.ordinal)
|
||
)).scalars())
|
||
|
||
return {
|
||
"document": DocumentOut.model_validate(doc).model_dump(mode="json"),
|
||
"chunks": [
|
||
{
|
||
"id": c.id,
|
||
"ordinal": c.ordinal,
|
||
"content": c.content[:_CHUNK_PREVIEW],
|
||
"truncated": len(c.content) > _CHUNK_PREVIEW,
|
||
"chars": len(c.content),
|
||
"token_len": c.token_len,
|
||
# 哪条向量是谁建的。一份文档里混着两个模型建的向量是真会发生的
|
||
# （换模型之后只重建了一部分），而那会让检索悄悄退回关键词
|
||
"embed_model": c.embed_model or "",
|
||
"has_vector": c.embedding is not None,
|
||
}
|
||
for c in rows
|
||
],
|
||
}
|
||
|
||
|
||
@kb_router.delete("/documents/{doc_id}", status_code=204)
|
||
async def delete_document(doc_id: str, session: AsyncSession = Depends(get_session)) -> None:
|
||
if not await kb.delete_document(session, doc_id):
|
||
raise HTTPException(404, "文档不存在")
|
||
|
||
|
||
# --------------------------------------------------------------------------
|
||
# Skill（方法论）
|
||
# --------------------------------------------------------------------------
|
||
|
||
skills_router = APIRouter(prefix="/api/skills", tags=["knowledge"])
|
||
|
||
|
||
class SkillIn(BaseModel):
|
||
name: str = Field(min_length=1, max_length=100)
|
||
description: str = ""
|
||
instructions: str = ""
|
||
examples: list[dict[str, Any]] = Field(default_factory=list)
|
||
suggested_tools: list[str] = Field(default_factory=list)
|
||
tags: list[str] = Field(default_factory=list)
|
||
enabled: bool = True
|
||
|
||
|
||
class SkillOut(SkillIn):
|
||
id: str
|
||
created_at: Any = None
|
||
|
||
model_config = {"from_attributes": True}
|
||
|
||
|
||
@skills_router.get("", response_model=list[SkillOut])
|
||
async def list_skills(session: AsyncSession = Depends(get_session)) -> list[Skill]:
|
||
return list((await session.execute(select(Skill).order_by(Skill.created_at.desc()))).scalars())
|
||
|
||
|
||
@skills_router.post("", response_model=SkillOut, status_code=201)
|
||
async def create_skill(payload: SkillIn, session: AsyncSession = Depends(get_session)) -> Skill:
|
||
if (await session.execute(select(Skill).where(Skill.name == payload.name))).scalar_one_or_none():
|
||
raise HTTPException(409, "同名 Skill 已存在")
|
||
row = Skill(**payload.model_dump())
|
||
session.add(row)
|
||
await session.commit()
|
||
await session.refresh(row)
|
||
return row
|
||
|
||
|
||
@skills_router.patch("/{skill_id}", response_model=SkillOut)
|
||
async def update_skill(
|
||
skill_id: str, payload: SkillIn, session: AsyncSession = Depends(get_session)
|
||
) -> Skill:
|
||
row = await session.get(Skill, skill_id)
|
||
if not row:
|
||
raise HTTPException(404, "Skill 不存在")
|
||
for key, value in payload.model_dump().items():
|
||
setattr(row, key, value)
|
||
await session.commit()
|
||
await session.refresh(row)
|
||
return row
|
||
|
||
|
||
@skills_router.delete("/{skill_id}", status_code=204)
|
||
async def delete_skill(skill_id: str, session: AsyncSession = Depends(get_session)) -> None:
|
||
row = await session.get(Skill, skill_id)
|
||
if not row:
|
||
raise HTTPException(404, "Skill 不存在")
|
||
await session.delete(row)
|
||
await session.commit()
|