首页 \ 问答 \ Web App架构问题(Web App architecture questions)

Web App架构问题(Web App architecture questions)

 背景：我是一位在.Net平台上工作的中级Web应用程序开发人员。 我的大部分工作都是由我的同事或上级为我定义的，我遵循指示并完成工作完全没有问题。  
 手头的任务：我最近被一位老朋友要求重新开始他的网络应用程序。 他的应用程序非常陈旧，而且他总是被打破。 有问题的应用程序是库存/ CRM应用程序，目前每个客户都需要新安装应用程序（通常通过在同一服务器上的不同域上部署并指向新数据库来完成）。  
 目前，如果任何客户想要对表单进行任何修改（例如附加字段，新功能等），我的朋友就会将这些字段手动添加到表单，脚本，数据库等。因此，此应用程序的所有安装都是唯一的。 没有一个单一的源代码库，没有一个单一版本的这个应用程序。 一般而言，新功能会加入到其他网站中，但仍按个别网站逐个网站完成。  
 我将在一个非常模块化的基础上接近这一点。 最初，我将编写一个模块来查询外部Web服务的某些数据，并显示和存储它，并定期自动更新它。 下一个模块可能用于存储和显示库存数据。 这种方式我想随着时间的推移重复他的应用程序的当前功能集100％，但增量。  
 百万美元的问题  
 
  我想让应用具有用户可配置的表单字段。 用户应该能够进入管理页面，创建特定类别的新表单页面，然后指定他想要的字段。 他可以说'创建一个名为Item＃的新文本字段并将其设置为需求'，并将其存储到某个位置。所有表单都将根据用户配置的内容动态呈现给用户。这是一种很好的方式不知道客户可能想要什么形式的问题，从而能够存储和显示任何形式的表单数据？我应该在这里遵循什么样的设计模式？ 
 我对asp.net和.net框架一般都很熟悉，并且对javascript，html，silverlight，jquery，c＃等具有相当的知识。我可以以一种很好的方式绕过web应用程序，但我不确定是什么我应该用这种框架或技术来完成这项任务。 ASP.net 3.5 webforms会成为一种方式吗？ 或者我应该看看ASP.NET MVC？ 我是否使用jquery和ajax完成前端和后端的解耦？ 或者将一个正常的asp.net页面与抛出的一个Ajax抛出的代码隐藏在一起是当天的顺序？  
 
 只是在开始之前寻找一般建议。  
 我目前正在考虑使用ASP.NET 3.5 webforms，用于客户端动画，UI，操作和数据验证的jQuery，以及用于后端的sqlserver + .net或wcf webservice。  
 你的建议一如既往地受到重视。 

Background: I am an intermediate web app developer working on the .Net Platform. Most of my work has been defined pretty well for me by my peers or superiors and I have no problem following instructions and getting the job done. 
The task at hand: I was recently asked by an old friend to redo his web app from scratch. His app is extremely antiquated and he is getting overwhelmed by it breaking all the time. The app in question is an inventory / CRM application and currently each customer requires a new install of the app (usually accomplished by deploying it on a different domain on the same server and pointing to a new database).  
Currently if any client wants any modifications to the forms such as additional fields, new features, etc my friend goes in and manually adds those fields to the forms, scripts, database etc. As a result all installs of this application are unique. There is no one singular source repository and no one single version of this app. Generally new features are overtime rolled into the other sites, but still this is done on an individual site by site basis.  
I will be approaching this on a very modular basis. Initially I will be coding a module that will query an external web service for some data, display and store it, and periodically update it automatically. The next module will likely be for storing and displaying inventory data. This way I want to over time duplicate the current feature set of his app 100% but do it incrementally. 
The Million Dollar Questions 
 
 I want to make the app have user configurable form fields. The user should be able to go to an admin page, create a new forms page of a certain category, and then specify what fields he wants in there. He could say 'create a new text field called Item # and make it a requirement" and that will get stored somewhere. All forms will be dynamically rendered to screen based on what the user has configured. Is this a good way to go about the problem of having no idea what a customer could want in a form? and thus be able to store and display form data of any sort ? What sort of design pattern should I follow here?
 I am familiar with asp.net and the .net framework in general and have decent knowledge of javascript, html, silverlight, jquery, c# etc etc. I can work my way around web apps in a good way, but I am not sure what sort of framework or tech I should use to accomplish this task. Would ASP.net 3.5 webforms be the way to go? or should I look into ASP.NET MVC? Do I use jquery and ajax for complete decoupling of frontend and backend ? or will a normal asp.net page with some spattering of ajax thrown in working with a codebehind be the order of the day? 
 
Just looking for general advice before I start. 
I am currently thinking of using ASP.NET 3.5 webforms, jquery for clientside animation, ui, manipulation and data validation, and sqlserver + a .net or wcf webservice for backend. 
Your advice is much appreciated as always. 

原文：https://stackoverflow.com/questions/1298373

更新时间：2022-04-20 16:04

最满意答案

 channel_swap将RGB转换为BGR ，如果您使用基于[1]中的注释的参考图像网络模型，则显然这是必要的。 在你的情况下，图像是灰度的，所以你可能没有三个通道。 您可能需要将其设置为（0，0，0），但即使这可能也无济于事（我不确定channel_swap的确切实施情况）。 如果这没有帮助，最简单的解决方案可能是通过将每个像素分成具有相同值的三个值（RGB）来预处理数据。 之后，您可能会完全删除channel_swap ，因为您的频道具有相同的值，并且交换它们是不可操作的。  
 意思是将从输入数据中减去将其居中。 （请记住，神经网络需要数据具有零均值，而输入图像通常具有正的均值，因此需要减法）。 您减去的平均值应与用于训练的平均值相同，因此使用与该模型相关的文件中的平均值是正确的。 不过，我不确定你是否应该调用.mean(1) - 你是否从某个例子中得到了这条线？ 如果是的话，那很可能是正确的做法。  
 raw_scale是您的输入数据的比例。 模型期望像素被归一化，所以如果你的输入数据的值在0到255之间，那么raw_scale设置为255是正确的。 如果您的数据的值介于0和1之间，则应将raw_scale设置为1。  
 最后，根据我对[2]中的评论的理解，您不需要提供image_dims  
 [1] https://github.com/BVLC/caffe/blob/master/python/caffe/io.py#L204  
 [2] https://github.com/BVLC/caffe/blob/master/python/caffe/classifier.py#L18 

The channel_swap is to reverse RGB into BGR, which is apparently necessary if you use a reference image net model, based on a comment in [1]. In your case the images are greyscale, so you probably do not have three channels. You might need to set it to (0, 0, 0), but even that might not help (I am unsure on the exact implementation of channel_swap). If that does not help, the simplest solution might be to preprocess you data by splitting every pixel into three values (RGB) with equal values. After that you might drop channel_swap altogether, because your channels have the same value, and swapping them is a no-op. 
Mean is what will be subtracted from your input data to center it. (Remember that neural networks need the data to have zero mean, while the input images usually have positive mean, hence the need of the subtraction). The mean you subtract should be the same that was used for training, so using mean from the file associated with the model is correct. I am not sure, however, on whether you should call .mean(1) on it -- did you get that line from some example? If yes, then it is most likely the correct thing to do. 
raw_scale is a scale of your input data. The model expects pixels to be normalized, so if your input data has values between 0 and 255, then raw_scale set to 255 is correct. If your data has values between 0 and 1, then raw_scale should be set to 1. 
Finally, based on my understanding of the comment in [2] you do not need to provide image_dims 
[1] https://github.com/BVLC/caffe/blob/master/python/caffe/io.py#L204 
[2] https://github.com/BVLC/caffe/blob/master/python/caffe/classifier.py#L18

Web App架构问题(Web App architecture questions)

最满意答案

相关问答

使用bvlc_googlenet作为数字中的预训练模型 - 错误(Using bvlc_googlenet as pretrained model in digits - errors)[2023-10-13]

可以直接从gpu给Caffe或Caffe2输入数据吗？(Can Caffe or Caffe2 be given input data directly from gpu?)[2023-10-13]

Caffe总是在Python中给出相同的预测，但训练准确性很好(Caffe always gives the same prediction in Python, but training accuracy is good)[2024-02-02]

如何保存预训练caffe模型的子集？(How a subset of pretrained caffe model can be saved?)[2022-04-20]

Caffe，如何从预训练网络预测(Caffe, how to predict from a pretrained net)[2023-09-01]

使用keras通过预训练的VGG实现感知损失(Implement perceptual loss with pretrained VGG using keras)[2023-06-07]

Caffe预训练模型具有更大的输入图像尺寸(Caffe pretrained models with larger input image sizes)[2022-07-26]

Caffe：尽可能简单地预测快照中的图像(Caffe: Predict an image from snapshot as simple as possible)[2024-02-10]

Caffe将AlexNet的PreTrained Weights复制到具有两个AlexNets的自定义网络(Caffe Copy PreTrained Weights of AlexNet to custom network that has Two AlexNets)[2022-04-18]

相关文章

最新问答