MDI without scrollbar

本文介绍了一种使用 Delphi 实现 MDI 应用中动态隐藏滚动条的方法。通过 API 级别操作,在 WM_NCCALCSIZE 消息中去除滚动条样式,从而实现滚动条的隐藏。


Etienne,

this has no effect since the scrollbars do not belong to the MDI frame
window itself, they belong to the client window, which is not a Delphi
form. Which means one has to attack the problem on the API level.
Since this question has come up so frequently in recent days i have
modified a sample based on the stock MDI project to include this
feature. The salient parts are quoted below.


Open the main forms unit in the IDE.
If you don't have a handler for the OnCreate event, add one. In the
handler you do this:


  If ClientHandle <> 0 Then Begin
    If GetWindowLong( ClientHandle, GWL_USERDATA ) <> 0 Then
      Exit; // cannot subclass client window, userdata already in use
    SetWindowLong( ClientHandle, GWL_USERDATA,
                   SetWindowLong( ClientHandle, GWL_WNDPROC,
                                  integer( @ClientWindowProc)));
  End;


Add a new standalone function to the unit, it has to go above the
FormCreate method since it is referenced in the statement above.


Function ClientWindowProc( wnd: HWND; msg: Cardinal; wparam,
                           lparam: Integer ): Integer; stdcall;
Var
  f: Pointer;
Begin
  f:= Pointer( GetWindowLong( wnd, GWL_USERDATA ));
  Case msg of
    WM_NCCALCSIZE: Begin
        If ( GetWindowLong( wnd, GWL_STYLE ) and
             (WS_HSCROLL or WS_VSCROLL)) <> 0
        Then
          SetWindowLong( wnd, GWL_STYLE,
                         GetWindowLong( wnd, GWL_STYLE )
                         and not (WS_HSCROLL or WS_VSCROLL));
      End;
  End;
  Result := CallWindowProc( f, wnd, msg, wparam, lparam );
End;


I clipped this code from a larger project, so lets hope i did not
create errors in the process. What this code does is to subclass the
client window the API way. It stores the old window function into the
GWL_USERDATA field of the window structure since it is needed in the
replacement window function, all messages need to be passed on to the
old window function. There is only one message of interest in this case
(the use of a Case results from the larger project, which handles more
than this message): WM_NCCALCSIZE. The window gets this message when
Windows tries to hide or show the scrollbars, among other cases. And it
arrives *before* there is any painting of the scrollbar. So we can
check if the window is going to sprout scrollbars and simply remove the
scrollbar styles again.


For the purists: there is no need to undo the subclassing before the
form is destroyed since the client window is destroyed before the form
object.

 

内容概要:本文介绍了基于贝叶斯优化的CNN-LSTM混合神经网络在时间序列预测中的应用,并提供了完整的Matlab代码实现。该模型结合了卷积神经网络(CNN)在特征提取方面的优势与长短期记忆网络(LSTM)在处理时序依赖问题上的强大能力,形成一种高效的混合预测架构。通过贝叶斯优化算法自动调参,提升了模型的预测精度与泛化能力,适用于风电、光伏、负荷、交通流等多种复杂非线性系统的预测任务。文中还展示了模型训练流程、参数优化机制及实际预测效果分析,突出其在科研与工程应用中的实用性。; 适合人群:具备一定机器学习基基于贝叶斯优化CNN-LSTM混合神经网络预测(Matlab代码实现)础和Matlab编程经验的高校研究生、科研人员及从事预测建模的工程技术人员,尤其适合关注深度学习与智能优化算法结合应用的研究者。; 使用场景及目标:①解决各类时间序列预测问题,如能源出力预测、电力负荷预测、环境数据预测等;②学习如何将CNN-LSTM模型与贝叶斯优化相结合,提升模型性能;③掌握Matlab环境下深度学习模型搭建与超参数自动优化的技术路线。; 阅读建议:建议读者结合提供的Matlab代码进行实践操作,重点关注贝叶斯优化模块与混合神经网络结构的设计逻辑,通过调整数据集和参数加深对模型工作机制的理解,同时可将其框架迁移至其他预测场景中验证效果。
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值