完善部分平台治理功能，及修改看板缺陷

backend/app/api/v1/endpoints/platform.py

@@ -1,4 +1,4 @@
-from __future__ import annotations
+from __future__ import annotations
 
 import json
 import uuid
@@ -442,16 +442,14 @@ async def dashboard_stats() -> dict[str, Any]:
 
     running_statuses = {"syncing", "queued", "running"}
     running_ft = [t for t in tasks if t.get("status") in running_statuses]
-    failed_ft = [t for t in tasks if t.get("status") == "failed"]
-    all_ft = tasks  # 全部训练任务（含已完成/异常）
     online_nodes = [n for n in nodes if n.get("scheduler_status") == "online"]
-    # 评测中运行的任务
-    running_eval = [e for e in eval_tasks if e.get("status") in running_statuses]
-    # 数据处理中运行的任务
+    # 评测中运行的任务数
+    eval_running = 0
     try:
-        dp_running = int(dp_store.list_tasks(page=1, page_size=1, status="running").get("total", 0))
+        eval_tasks = store.eval_tasks()
+        eval_running = len([e for e in eval_tasks if e.get("status") in running_statuses])
     except Exception:
-        dp_running = 0
+        eval_running = 0
 
     # 近 7 天训练统计（按创建日期分桶）
     now = datetime.now(timezone.utc)
@@ -472,45 +470,30 @@ async def dashboard_stats() -> dict[str, Any]:
             }
         )
 
-    # 服务状态 —— 通过对应接口连通性判断是否正常
+    # 服务状态 —— 每个服务的"实例数"含义：
+    #   模型训练 → 训练任务总数
+    #   模型评测 → 评测任务总数
+    #   模型推理 → 推理/对比任务实例数
+    #   模型管理 → 基座模型注册总数
+    #   数据集管理 → 数据集总数
+    #   数据处理 → 数据处理任务总数
+    #   数据类型转换 → 数据转换任务总数
     service_checks = [
-        ("模型训练", "/fine-tune", "模型训练"),
-        ("模型评测", "/model-eval", "模型评测"),
-        ("模型推理", "/model-inference", "模型推理"),
-        ("模型管理", "/model-manage", "模型管理"),
-        ("数据集管理", "/dataset-manage", "数据集管理"),
-        ("数据处理", "/data-process", "数据处理"),
-        ("数据类型转换", "/data-convert", "数据类型转换"),
+        ("模型训练", "fine-tune", len(tasks)),
+        ("模型评测", "model-eval", len(eval_tasks)),
+        ("模型推理", "model-inference", len(store.compare_tasks())),
+        ("模型管理", "model-manage", len(store.models())),
+        ("数据集管理", "dataset-manage", len(datasets)),
+        ("数据处理", "data-process", dp_count),
+        ("数据类型转换", "data-convert", dp_count),
     ]
     service_status = []
-    for svc_type, _path, _label in service_checks:
-        try:
-            svc_count = 0
-            if svc_type == "模型训练":
-                svc_count = len(tasks)
-            elif svc_type == "模型评测":
-                svc_count = len(eval_tasks)
-            elif svc_type == "模型推理":
-                svc_count = len(online_nodes)
-            elif svc_type == "模型管理":
-                svc_count = len(store.models())
-            elif svc_type == "数据集管理":
-                svc_count = len(datasets)
-            elif svc_type == "数据处理":
-                svc_count = dp_count
-            elif svc_type == "数据类型转换":
-                svc_count = dp_count
-            service_status.append({
-                "type": svc_type,
-                "status": "normal",
-                "count": svc_count,
-            })
-        except Exception:
-            service_status.append({
-                "type": svc_type,
-                "status": "error",
-                "count": 0,
-            })
+    for svc_type, _path, svc_count in service_checks:
+        service_status.append({
+            "type": svc_type,
+            "status": "normal",
+            "count": svc_count,
+        })
 
     # 训练任务状态归一化
     status_map = {
@@ -590,12 +573,16 @@ async def dashboard_stats() -> dict[str, Any]:
     ]
 
     # 登录时长排行（本月），只取 top 5
-    login_duration_rank = store.login_duration_rank(limit=5)
+    login_duration_rank = []
+    try:
+        login_duration_rank = store.login_duration_rank(limit=5)
+    except Exception:
+        pass
 
     return ok(
         {
             "online_services": sum(s["count"] for s in service_status),
-            "running_tasks": len(running_ft) + len(running_eval) + dp_running,
+            "running_tasks": len(running_ft) + eval_running,
             "pending_alerts": 0,
             "training_7d": training_7d,
             "service_status": service_status,
@@ -656,6 +643,29 @@ async def reset_user_password(
         raise fail(400, str(exc))
 
 
+@router.post("/users/me/password")
+async def change_my_password(
+    payload: dict[str, Any] = Body(...),
+    current_user: dict = Depends(get_current_user),
+) -> dict[str, Any]:
+    """用户自行修改密码：验证旧密码后设置新密码。"""
+    old_password = payload.get("old_password") or ""
+    new_password = payload.get("new_password") or ""
+    if not old_password or not new_password:
+        raise fail(400, "old_password and new_password are required")
+    if len(new_password) < 6:
+        raise fail(400, "new password must be at least 6 characters")
+    try:
+        success = get_platform_store().change_password(
+            current_user["id"], old_password, new_password
+        )
+    except KeyError:
+        raise fail(404, "user not found")
+    if not success:
+        raise fail(400, "old password is incorrect")
+    return ok({"changed": True})
+
+
 @router.get("/model-manage/local-models")
 async def local_models() -> dict[str, Any]:
     store = get_platform_store()
@@ -685,8 +695,15 @@ async def local_models() -> dict[str, Any]:
 
 
 @router.get("/model-manage/trained-models")
-async def trained_models() -> dict[str, Any]:
-    return ok({"models": get_platform_store().trained_models()})
+async def trained_models(current_user: dict = Depends(get_current_user)) -> dict[str, Any]:
+    all_models = get_platform_store().trained_models()
+    if is_admin(current_user):
+        return ok({"models": all_models})
+    # 普通用户只能看到自己创建的 + ACL 授权的
+    user_id = current_user.get("id")
+    accessible = set(filter_accessible_resource_ids("trained_model", [m["id"] for m in all_models], current_user))
+    result = [m for m in all_models if m.get("created_by") == user_id or m["id"] in accessible]
+    return ok({"models": result})
 
 
 @router.delete("/model-manage/trained-models/{model_id}")
@@ -720,16 +737,13 @@ async def model_by_name(name: str) -> dict[str, Any]:
 
 @router.get("/model-manage")
 async def model_list(current_user: dict = Depends(get_current_user)) -> dict[str, Any]:
-    models = get_platform_store().models()
-    if current_user.get("role") == "admin" or current_user.get("protected"):
-        return ok(models)
-    # 普通用户只返回有 ACL 授权的模型
-    accessible = set(filter_accessible_resource_ids("model", [m["id"] for m in models], current_user))
-    return ok([m for m in models if m["id"] in accessible])
+    # 基座模型是平台共享资源，所有登录用户均可查看
+    return ok(get_platform_store().models())
 
 
 @router.post("/model-manage")
-async def create_model(payload: dict[str, Any] = Body(...)) -> dict[str, Any]:
+async def create_model(payload: dict[str, Any] = Body(...), current_user: dict = Depends(get_current_user)) -> dict[str, Any]:
+    payload.setdefault("created_by", current_user.get("id"))
     try:
         return ok(get_platform_store().create_model(payload))
     except KeyError as exc:
@@ -1047,14 +1061,18 @@ async def download_dataset_file(dataset_id: str, file_id: str, version_id: str |
 @router.get("/dataset-manage")
 async def dataset_list(current_user: dict = Depends(get_current_user)) -> dict[str, Any]:
     datasets = get_platform_store().datasets()
-    if current_user.get("role") == "admin" or current_user.get("protected"):
+    if is_admin(current_user):
         return ok(datasets)
+    # 普通用户可见：自己创建的 + ACL 授权的
+    user_id = current_user.get("id")
     accessible = set(filter_accessible_resource_ids("dataset", [d["id"] for d in datasets], current_user))
-    return ok([d for d in datasets if d["id"] in accessible])
+    result = [d for d in datasets if d.get("created_by") == user_id or d["id"] in accessible]
+    return ok(result)
 
 
 @router.post("/dataset-manage")
-async def create_dataset(payload: dict[str, Any] = Body(...)) -> dict[str, Any]:
+async def create_dataset(payload: dict[str, Any] = Body(...), current_user: dict = Depends(get_current_user)) -> dict[str, Any]:
+    payload.setdefault("created_by", current_user.get("id"))
     dataset = get_platform_store().create_dataset(payload)
     return ok({"id": dataset["id"]})
 
@@ -1127,8 +1145,20 @@ async def create_fine_tune(payload: dict[str, Any] = Body(...)) -> dict[str, Any
 
 
 @router.post("/fine-tune/start")
-async def start_fine_tune(payload: dict[str, Any] = Body(...)) -> dict[str, Any]:
+async def start_fine_tune(
+    payload: dict[str, Any] = Body(...),
+    current_user: dict = Depends(get_current_user),
+) -> dict[str, Any]:
     store = get_platform_store()
+    # GPU 权限校验：普通用户只能使用被分配的 GPU
+    if not is_admin(current_user):
+        node_id = payload.get("compute_node_id") or payload.get("node_id")
+        gpu_indices = payload.get("gpus") or []
+        if node_id and gpu_indices:
+            if not store.check_gpu_access(current_user["id"], node_id, gpu_indices):
+                raise fail(403, "无权使用所选 GPU，请联系管理员分配")
+    # 记录创建者
+    payload.setdefault("created_by", current_user.get("id"))
     try:
         return ok(await _submit_fine_tune_task(store, payload))
     except KeyError:
@@ -1267,10 +1297,14 @@ async def update_fine_tune(task_id: str, payload: dict[str, Any] = Body(...)) ->
 
 
 @router.post("/fine-tune/stop/{task_id}")
-async def stop_fine_tune(task_id: str) -> dict[str, Any]:
+async def stop_fine_tune(task_id: str, current_user: dict = Depends(get_current_user)) -> dict[str, Any]:
     store = get_platform_store()
     try:
         task = store.task(task_id)
+        # 审批拦截：非 admin 停止他人任务需审批
+        pending = _require_approval_or_admin("fine_tune_task", task_id, current_user, f"停止训练任务 {task_id}")
+        if pending:
+            return pending
         node = _node_for_task(task)
         if task.get("compute_job_id") and node and get_settings().compute_mode != "simulator":
             job = await ComputeNodeClient(node["api_base_url"]).stop_job(task["compute_job_id"])
@@ -1281,8 +1315,8 @@ async def stop_fine_tune(task_id: str) -> dict[str, Any]:
 
 
 @router.post("/fine-tune/{task_id}/stop")
-async def stop_fine_tune_alt(task_id: str) -> dict[str, Any]:
-    return await stop_fine_tune(task_id)
+async def stop_fine_tune_alt(task_id: str, current_user: dict = Depends(get_current_user)) -> dict[str, Any]:
+    return await stop_fine_tune(task_id, current_user)
 
 
 @router.post("/fine-tune/{task_id}/retry")

backend/app/api/v1/router.py

@@ -1,4 +1,4 @@
-from fastapi import APIRouter
+from fastapi import APIRouter
 
 from app.api.v1.endpoints.data_process import router as data_process_router
 from app.api.v1.endpoints.platform import router as platform_router
@@ -9,6 +9,7 @@ from app.modules.approval.router import router as approval_router
 from app.modules.system.router import router as system_router
 from app.modules.retention.router import router as retention_router
 from app.modules.resource.router import router as resource_router
+from app.modules.gpu.router import router as gpu_router
 from app.modules.data_convert.router import router as data_convert_router
 
 api_router = APIRouter()
@@ -21,4 +22,5 @@ api_router.include_router(project_router, tags=["project"])
 api_router.include_router(approval_router, tags=["approval"])
 api_router.include_router(retention_router, tags=["retention"])
 api_router.include_router(resource_router, tags=["resource"])
+api_router.include_router(gpu_router, tags=["gpu-assignment"])
 api_router.include_router(data_convert_router, tags=["data-convert"])