fix: complete admin knowledge audit improvements

ai_knowledge_base_v2/apps/backend/app/api/admin_agent_records.py

@@ -1,9 +1,10 @@
 from __future__ import annotations
 
 import json
+from datetime import datetime
 
 from fastapi import APIRouter, Depends, HTTPException, Query, status
-from sqlalchemy import select
+from sqlalchemy import delete, func, select
 from sqlalchemy.orm import Session
 
 from app.core.database import get_db
@@ -18,6 +19,8 @@ from app.models.knowledge import (
 )
 from app.schemas.knowledge import AttentionUpdateRequest
 from app.services.admin_service import OperationLogService
+from app.models.logs import LogRetentionPolicy
+from app.services.maintenance_service import MaintenanceService
 
 router = APIRouter()
 
@@ -29,10 +32,12 @@ def retrieval_logs(
     knowledgeCalled: bool | None = Query(default=None),
     statusValue: str = Query(default="", alias="status"),
     attentionCreated: bool | None = Query(default=None),
+    page: int = Query(default=1, ge=1),
+    pageSize: int = Query(default=50, ge=1, le=100),
     db: Session = Depends(get_db),
     current_admin: Admin = Depends(get_current_admin),
 ) -> dict:
-    query = select(KnowledgeRetrievalLog).order_by(KnowledgeRetrievalLog.id.desc()).limit(300)
+    query = select(KnowledgeRetrievalLog).order_by(KnowledgeRetrievalLog.id.desc())
     if sessionId is not None:
         query = query.where(KnowledgeRetrievalLog.session_id == sessionId)
     if knowledgeId is not None:
@@ -43,9 +48,55 @@ def retrieval_logs(
         query = query.where(KnowledgeRetrievalLog.status == statusValue)
     if attentionCreated is not None:
         query = query.where(KnowledgeRetrievalLog.attention_created == (1 if attentionCreated else 0))
+    query = query.offset((page - 1) * pageSize).limit(pageSize)
     return api_success([_retrieval_dict(item, detail=False) for item in db.scalars(query).all()])
 
 
+@router.post("/retrieval-log/cleanup/estimate")
+def estimate_cleanup(payload: dict, db: Session = Depends(get_db), current_admin: Admin = Depends(get_current_admin)) -> dict:
+    before = _cleanup_before(payload)
+    count = db.scalar(select(func.count()).select_from(KnowledgeRetrievalLog).where(KnowledgeRetrievalLog.created_at < before)) or 0
+    return api_success({"before": before, "estimatedCount": count})
+
+
+@router.post("/retrieval-log/cleanup")
+def cleanup_logs(payload: dict, db: Session = Depends(get_db), current_admin: Admin = Depends(get_current_admin)) -> dict:
+    before = _cleanup_before(payload)
+    total = MaintenanceService.delete_retrieval_logs(db, before)
+    OperationLogService.write(db, admin_id=current_admin.id, module="retrieval_log", action=f"cleanup:{total}", target_id=None)
+    db.commit()
+    return api_success({"deleted": total, "before": before})
+
+
+@router.get("/retrieval-log/retention")
+def get_retention(db: Session = Depends(get_db), current_admin: Admin = Depends(get_current_admin)) -> dict:
+    policy = db.scalar(select(LogRetentionPolicy).order_by(LogRetentionPolicy.id).limit(1))
+    return api_success({"enabled": bool(policy and policy.enabled), "retentionDays": policy.retention_days if policy else None, "lastRunAt": policy.last_run_at if policy else None, "lastResult": policy.last_result if policy else None})
+
+
+@router.put("/retrieval-log/retention")
+def set_retention(payload: dict, db: Session = Depends(get_db), current_admin: Admin = Depends(get_current_admin)) -> dict:
+    days = payload.get("retentionDays")
+    if days is not None and (not isinstance(days, int) or days < 1):
+        raise HTTPException(status_code=400, detail="保留天数必须为正整数，永久保留请传 null")
+    policy = db.scalar(select(LogRetentionPolicy).order_by(LogRetentionPolicy.id).limit(1)) or LogRetentionPolicy()
+    policy.retention_days, policy.enabled, policy.updated_by = days, 0 if days is None else 1, current_admin.id
+    db.add(policy)
+    OperationLogService.write(db, admin_id=current_admin.id, module="retrieval_log", action="retention_update", target_id=None)
+    db.commit()
+    return api_success({"enabled": bool(policy.enabled), "retentionDays": policy.retention_days})
+
+
+def _cleanup_before(payload: dict) -> datetime:
+    raw = str(payload.get("before") or "").strip()
+    if not raw:
+        raise HTTPException(status_code=400, detail="必须指定清理日期")
+    try:
+        return datetime.fromisoformat(raw.replace("Z", "+00:00")).replace(tzinfo=None)
+    except ValueError as exc:
+        raise HTTPException(status_code=400, detail="清理日期格式错误") from exc
+
+
 @router.get("/retrieval-log/{log_id}")
 def retrieval_detail(
     log_id: int,

ai_knowledge_base_v2/apps/backend/app/api/admin_dashboard.py

@@ -1,8 +1,11 @@
 from __future__ import annotations
 
-from datetime import datetime
+from datetime import datetime, timedelta
 
 from fastapi import APIRouter, Depends, Query
+import json
+from pathlib import Path
+from sqlalchemy import select, text
 from sqlalchemy.orm import Session
 
 from app.core.database import get_db
@@ -11,6 +14,8 @@ from app.core.responses import api_success
 from app.models.admin import Admin
 from app.schemas.admin import DashboardStats
 from app.services.admin_service import AdminDashboardService
+from app.models.logs import StorageSnapshot
+from app.services.redis_client import get_sync_redis_client
 
 router = APIRouter()
 
@@ -30,3 +35,53 @@ def dashboard(
         end_dt = None
     stats = AdminDashboardService.stats(db, start_dt, end_dt)
     return api_success(DashboardStats.model_validate(stats).model_dump())
+
+
+@router.get("/dashboard/storage")
+def storage_stats(
+    refresh: bool = Query(default=False),
+    db: Session = Depends(get_db),
+    current_admin: Admin = Depends(get_current_admin),
+) -> dict:
+    latest = db.scalar(select(StorageSnapshot).where(StorageSnapshot.scope == "project").order_by(StorageSnapshot.id.desc()).limit(1))
+    if latest is not None and not refresh:
+        return api_success(_snapshot_dict(latest, db))
+    detail: dict = {"database": {"status": "unavailable", "bytes": None, "tables": []}, "redis": {"status": "unavailable", "bytes": None}, "files": {"status": "unavailable", "bytes": None}}
+    errors: list[str] = []
+    try:
+        if db.bind and db.bind.dialect.name == "mysql":
+            rows = db.execute(text("SELECT table_name, data_length + index_length AS bytes FROM information_schema.tables WHERE table_schema = DATABASE() ORDER BY bytes DESC")).all()
+            tables = [{"name": row[0], "bytes": int(row[1] or 0)} for row in rows]
+            detail["database"] = {"status": "ok", "bytes": sum(x["bytes"] for x in tables), "tables": tables}
+        else:
+            detail["database"] = {"status": "unavailable", "bytes": None, "tables": [], "message": "当前数据库不支持精确表占用统计"}
+    except Exception as exc:
+        errors.append(f"database: {exc}")
+    try:
+        redis = get_sync_redis_client()
+        if redis is not None:
+            info = redis.info("memory")
+            detail["redis"] = {"status": "ok", "bytes": int(info.get("used_memory", 0))}
+    except Exception as exc:
+        errors.append(f"redis: {exc}")
+    try:
+        root = Path(".uploads")
+        if root.exists():
+            detail["files"] = {"status": "ok", "bytes": sum(p.stat().st_size for p in root.rglob("*") if p.is_file())}
+        else:
+            detail["files"] = {"status": "unavailable", "bytes": None, "message": "项目未配置本地附件目录"}
+    except Exception as exc:
+        errors.append(f"files: {exc}")
+    known = [x.get("bytes") for x in detail.values() if x.get("status") == "ok"]
+    snapshot = StorageSnapshot(scope="project", total_bytes=sum(known) if known else None, detail_json=json.dumps(detail, ensure_ascii=False), status="partial" if errors or any(x.get("status") != "ok" for x in detail.values()) else "ok", error_message="; ".join(errors) or None)
+    db.add(snapshot); db.commit(); db.refresh(snapshot)
+    return api_success(_snapshot_dict(snapshot, db))
+
+
+def _snapshot_dict(item: StorageSnapshot, db: Session) -> dict:
+    with_growth = {"totalBytes": item.total_bytes, "detail": json.loads(item.detail_json), "status": item.status, "errorMessage": item.error_message, "createdAt": item.created_at}
+    # Growth is intentionally unknown until enough snapshots exist; never report a misleading zero.
+    for days, key in ((7, "growth7DaysBytes"), (30, "growth30DaysBytes")):
+        previous = db.scalar(select(StorageSnapshot).where(StorageSnapshot.scope == item.scope, StorageSnapshot.created_at <= item.created_at - timedelta(days=days)).order_by(StorageSnapshot.created_at.desc()).limit(1))
+        with_growth[key] = item.total_bytes - previous.total_bytes if previous and item.total_bytes is not None and previous.total_bytes is not None else None
+    return with_growth

ai_knowledge_base_v2/apps/backend/app/api/admin_knowledge.py

@@ -36,6 +36,31 @@ def list_knowledge(db: Session = Depends(get_db), current_admin: Admin = Depends
     return api_success([_knowledge_dict(item) for item in items])
 
 
+@router.post("/knowledge/resolve-node")
+def resolve_knowledge_node(
+    payload: dict,
+    db: Session = Depends(get_db),
+    current_admin: Admin = Depends(get_current_admin),
+) -> dict:
+    _require_super_admin(current_admin)
+    node_id = str(payload.get("nodeId") or "").strip()
+    if not node_id or len(node_id) > 100 or not node_id.replace("_", "").replace("-", "").isalnum():
+        raise HTTPException(status_code=400, detail="Node ID 格式错误")
+    existing = db.scalar(select(Knowledge).where(Knowledge.feishu_node_id == node_id))
+    if existing is not None:
+        raise HTTPException(status_code=409, detail=f"该 Node ID 已添加（知识库：{existing.name}）")
+    try:
+        info = FeishuKnowledgeService.get_node_info(node_id, db)
+    except Exception as exc:
+        raise HTTPException(status_code=502, detail=f"飞书节点信息获取失败：{exc}") from exc
+    space_id, title = str(info.get("space_id") or "").strip(), str(info.get("title") or "").strip()
+    if not space_id:
+        raise HTTPException(status_code=422, detail="飞书接口未返回 Space ID，请检查节点是否移动或当前应用权限")
+    if not title:
+        raise HTTPException(status_code=422, detail="飞书接口未返回文章标题，请检查文章是否已删除")
+    return api_success({"nodeId": node_id, "spaceId": space_id, "sourceTitle": title, "name": title, "remark": str(info.get("description") or ""), "objType": info.get("obj_type")})
+
+
 @router.post("/knowledge/import")
 def import_knowledge_from_space(
     payload: dict,
@@ -91,12 +116,7 @@ def create_knowledge(
     current_admin: Admin = Depends(get_current_admin),
 ) -> dict:
     _require_super_admin(current_admin)
-    existing = db.scalar(
-        select(Knowledge).where(
-            Knowledge.feishu_space_id == payload.feishuSpaceId,
-            Knowledge.feishu_node_id == payload.feishuNodeId,
-        )
-    )
+    existing = db.scalar(select(Knowledge).where(Knowledge.feishu_node_id == payload.feishuNodeId))
     if existing is not None:
         raise HTTPException(status_code=status.HTTP_409_CONFLICT, detail="该飞书文章已创建知识库")
     knowledge = Knowledge(