amb/amb_rag
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

细节优化

Browse Source
This commit is contained in:
amb
2026-09-02 18:32:56 +08:00
parent 8ec7856bfc
commit bfb7a1a8a2
7 changed files with 69 additions and 53 deletions
Download Patch File Download Diff File Expand all files Collapse all files
backend/app/api/categories.py
+35 -17
View File
@@ -52,10 +52,13 @@ def get_category_tree(
"""获取完整目录树（含文档数量）。"""
_check_kb_owner(kb_id, user, db)
# 获取所有分类
# 获取所有未删除的分类
stmt = (
select(DocumentCategory)
.where(DocumentCategory.knowledge_base_id == kb_id)
.where(
DocumentCategory.knowledge_base_id == kb_id,
DocumentCategory.deleted_at.is_(None),
)
.order_by(DocumentCategory.sort_order, DocumentCategory.name)
)
all_cats = list(db.scalars(stmt).all())
@@ -110,7 +113,10 @@ def list_categories_flat(
stmt = (
select(DocumentCategory)
.where(DocumentCategory.knowledge_base_id == kb_id)
.where(
DocumentCategory.knowledge_base_id == kb_id,
DocumentCategory.deleted_at.is_(None),
)
.order_by(DocumentCategory.path, DocumentCategory.sort_order)
)
cats = list(db.scalars(stmt).all())
@@ -236,32 +242,44 @@ def delete_category(
user: User = Depends(get_current_user),
db: Session = Depends(get_db),
) -> None:
"""删除分类（同时删除子分类，关联文档的 category_id 置 NULL）。"""
"""删除分类 → 进回收站（含全部子分类及其下文档，3 天后自动彻底删除）。"""
from datetime import datetime
from app.models.document import Document
_check_kb_owner(kb_id, user, db)
cat = db.get(DocumentCategory, cat_id)
if cat is None or cat.knowledge_base_id != kb_id:
if cat is None or cat.knowledge_base_id != kb_id or cat.deleted_at:
raise NotFoundError("分类不存在。")
# 删除所有子分类
now = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
# 收集本分类 + 所有存活的后代分类
target_ids = [cat_id]
if cat.path:
stmt = select(DocumentCategory).where(
DocumentCategory.knowledge_base_id == kb_id,
DocumentCategory.path.startswith(cat.path),
DocumentCategory.id != cat_id,
DocumentCategory.deleted_at.is_(None),
)
children = list(db.scalars(stmt).all())
for child in children:
db.delete(child)
for c in db.scalars(stmt).all():
if c.id != cat_id:
target_ids.append(c.id)
# 关联文档的 category_id 置 NULL
from app.models.document import Document
# 软删除分类树
for cid in target_ids:
c = db.get(DocumentCategory, cid)
c.deleted_at = now
doc_stmt = select(Document).where(Document.category_id == cat_id)
docs = list(db.scalars(doc_stmt).all())
for doc in docs:
doc.category_id = None
# 软删除目录下的文档（同一批次时间戳，便于整组恢复）
doc_stmt = select(Document).where(
Document.category_id.in_(target_ids),
Document.status != "DELETED",
)
for doc in db.scalars(doc_stmt).all():
doc.status = "DELETED"
doc.deleted_at = now
db.delete(cat)
db.commit()
backend/app/core/config.py
+2
View File
@@ -41,6 +41,8 @@ class Settings(BaseSettings):
# --- 数据保留 ---
# 访问日志保留天数，超过自动清理（0 = 永久保留，不推荐）
access_log_retention_days: int = 90
# 回收站保留天数，超期自动彻底删除（含物理文件）
recycle_bin_retention_days: int = 3
# --- CORS ---
frontend_origin: str = "http://localhost:5173"
backend/app/models/document.py
+5
View File
@@ -126,6 +126,11 @@ class Document(UUIDPrimaryKeyMixin, TimestampMixin, Base):
nullable=False,
comment="状态 (PENDING/PROCESSING/READY/FAILED/DELETED)",
)
deleted_at: Mapped[str | None] = mapped_column(
String(32),
nullable=True,
comment="进入回收站时间（NULL=未删除），3 天后自动清理",
)
error_code: Mapped[str | None] = mapped_column(
String(64),
nullable=True,
backend/app/models/document_category.py
+5
View File
@@ -52,6 +52,11 @@ class DocumentCategory(UUIDPrimaryKeyMixin, TimestampMixin, Base):
nullable=False,
comment="排序序号",
)
deleted_at: Mapped[str | None] = mapped_column(
String(32),
nullable=True,
comment="进入回收站时间（NULL=未删除），3 天后自动清理",
)
# 关系
knowledge_base = relationship("KnowledgeBase", back_populates="categories", lazy="selectin")
backend/app/models/knowledge_base.py
+5
View File
@@ -65,6 +65,11 @@ class KnowledgeBase(UUIDPrimaryKeyMixin, TimestampMixin, Base):
nullable=True,
comment="链接过期时间，NULL=长期有效",
)
deleted_at: Mapped[str | None] = mapped_column(
String(32),
nullable=True,
comment="进入回收站时间（NULL=未删除），3 天后自动清理",
)
# 关系
user = relationship("User", back_populates="knowledge_bases", lazy="selectin")
backend/app/services/doc_service.py
+5 -6
View File
@@ -1,6 +1,7 @@
"""文档服务：上传、删除、配额管理。"""
import hashlib
from datetime import datetime
from pathlib import Path
import filetype
@@ -153,16 +154,14 @@ class DocumentService:
return self._doc_repo.list_by_knowledge_base(kb_id, category_id=category_id, page=page, page_size=page_size)
def delete(self, doc_id: str, user: User) -> None:
"""删除文档 → 进回收站（3 天后自动彻底删除）。文件暂保留以支持恢复。"""
doc = self.get_or_404(doc_id, user)
file_size = doc.file_size
self._doc_repo.delete(doc)
self._delete_doc_files(doc)
# 回补配额
self._restore_quota(user, file_size)
doc.deleted_at = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
self._session.commit()
def _delete_doc_files(self, doc: Document) -> None:
"""物理删除文档的原始文件与 Markdown 文件（软删后调用）。"""
def purge_files(self, doc: Document) -> None:
"""物理删除文档的原始文件与 Markdown 文件（彻底删除时调用）。"""
from app.storage.local_storage import get_storage
storage = get_storage()
backend/app/services/kb_service.py
+12 -30
View File
@@ -1,5 +1,7 @@
"""知识库服务：CRUD + Token 管理 + 默认目录树。"""
from datetime import datetime
from app.core.errors import NotFoundError, PermissionDeniedError
from app.core.logging import get_logger
from app.models.document_category import DocumentCategory
@@ -33,6 +35,10 @@ class KnowledgeBaseService:
token_encrypted=token_encrypted,
token_hint=token_hint,
)
# AI 链接默认 30 分钟有效期
from datetime import datetime, timedelta
kb.token_expires_at = (datetime.now() + timedelta(minutes=30)).strftime("%Y-%m-%d %H:%M:%S")
self._seed_default_categories(kb.id)
self._session.commit()
return kb, token
@@ -40,7 +46,7 @@ class KnowledgeBaseService:
def _seed_default_categories(self, kb_id: str) -> None:
"""创建默认目录树结构。"""
default_tree = [
("01 公司层", ["公司基本信息", "经营理念", "四大价值", "A/M/B三态"]),
("01 公司层", ["公司基本信息", "经营理念", "四大价值"]),
("02 战略层", ["公司战略", "客户战略", "AI战略", "产品战略"]),
("03 部门层", ["企划", "技术", "交付", "市场"]),
("04 岗位/AI角色", []),
@@ -97,39 +103,14 @@ class KnowledgeBaseService:
return kb
def delete(self, kb_id: str, user: User) -> None:
"""删除知识库 → 进回收站（3 天后自动彻底删除）。"""
kb = self.get_or_404(kb_id, user)
self._kb_repo.delete(kb)
self._cleanup_kb_files(kb)
kb.deleted_at = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
self._session.commit()
def _cleanup_kb_files(self, kb: KnowledgeBase) -> None:
"""物理删除知识库下所有文档的原始文件与 Markdown 文件（软删后调用）。
文件删除失败不阻塞删除流程（记录日志，可由后续清理兜底）。
"""
from sqlalchemy import select
from app.models.document import Document
from app.storage.local_storage import get_storage
stmt = select(Document).where(Document.knowledge_base_id == kb.id)
docs = list(self._session.scalars(stmt).all())
storage = get_storage()
removed = 0
for doc in docs:
for key in (doc.storage_path, doc.markdown_path):
if not key:
continue
try:
storage.delete(key)
removed += 1
except Exception as exc: # noqa: BLE001
logger.warning("清理文件失败 key=%s: %s", key, exc)
if removed:
logger.info("KB %s 软删，已物理清理 %d 个文件", kb.id, removed)
def regenerate_token(self, kb_id: str, user: User) -> tuple[KnowledgeBase, str]:
"""重新生成 Token。旧链接立即失效。"""
"""重新生成 Token。旧链接立即失效，新链接默认 30 分钟有效期。"""
kb = self.get_or_404(kb_id, user)
token, token_hash, token_encrypted, token_hint = self._token_svc.create_token_pair()
self._kb_repo.update(
@@ -138,6 +119,7 @@ class KnowledgeBaseService:
token_encrypted=token_encrypted,
token_hint=token_hint,
)
kb.token_expires_at = (datetime.now() + timedelta(minutes=30)).strftime("%Y-%m-%d %H:%M:%S")
self._session.commit()
return kb, token
@@ -149,7 +131,7 @@ class KnowledgeBaseService:
def set_expiry(self, kb_id: str, user: User, expires_in_minutes: int | None) -> KnowledgeBase:
"""设置链接有效期。None = 长期有效；负数表示已过期（测试用）。"""
from datetime import datetime, timedelta
from datetime import timedelta
kb = self.get_or_404(kb_id, user)
if expires_in_minutes is None: