Building Web Based Temperature/Humidity Monitor App With Raspberry PI 3 and Azure IoT Hub

2. April 2017 07:39

ASP.NET MVC Azure Cloud IoT Projects Raspberry Pi Web Windows 

I was planning to build a web based temperature/humidity monitoring app with Raspberry PI 3 and Windows 10 IoT Core. I have playing around with Raspbian for a while now and this time I wanted to try something different. I have been using Python to write application for Raspberry PI, Windows IoT Core with C# will be something new for me and therefore, this time I choose to go with Windows 10 IoT Core.

Here is the list of things to get started:

  • Raspberry PI 3 with Windows 10 IoT Core
  • Adafruit BME280 Temperature Humidity Pressure sensor\
  • Jumper wires
  • Power Bank
  • Visual Studio 2015/2017
  • IoT Dashboard
  • Azure account

Setting up the PI with Windows 10 IoT Core is a very straight forward process. You can read about it more here. I recommend to connect the device to your Wi-Fi network. This will give you a benefit to move your device around. The problem for the first time users will be that the device will not auto-connect to the Wi-Fi network. To solve this problem, you need to use a wired LAN connection from the network or from the local computer. Once plugged in you will be able to see the device in the IoT dashboard from where you can go to Device Portal and connect to the Wi-Fi network. After the connection is successful, you can then remove the LAN cable. To follow the exact steps, head over to this documentation. This is how your device will get displayed in the IoT dashboard.

Getting things ready in Azure

You can set up things in the cloud from the Azure Portal as well as from the IoT Dashboard itself. From dashboard you can create and provisioned a new IoT Hub and register the device in that Hub. Here is how you can do that. First sign-in to your Azure account by clicking the Sign-In option in the left hand side bar. After the sign-in is successful, you will be prompted with the below screen where you can provision a new IoT Hub and add your current device in the hub.

I am skipping the step for setting the IoT Hub from the Azure Portal as there is a good documentation available at Azure Docs. Pushing data to the IoT Hub with the SDK is easy but you can't make use of the IoT Hub alone to read that data and do your work. Therefore, you also need to set up a Queue to read that data. As of now, you cannot create a Queue from the new Azure Portal instead you have to head towards the old portal at manage.windowsazure.com and create a queue from there. To create a queue, click the New>App Services>Service Bus>Queue>Quick Create. Fill in the details to create a new queue.

Make sure that the region you are selecting here for the queue, it should be the same as your IoT Hub region. If you select a different region than that of your IoT Hub, you will not be able to associate this queue with the hub endpoint.

In the Azure Portal, to associate this queue with the IoT Hub endpoint, click the Endpoints and then click Add to add the queue to the endpoint.

Setting up the circuit

Now as you have the device and the OS ready, you can now build the circuit. The BME280 sensor can be used with both I2C and SPI. For the project, I am using I2C. Below is the Fritzing diagram for your reference. You can download the Fritzing sketch file on Github repo.

Creating Device Application

I am using Visual Studio 2017 Enterprise Edition to build the device application and the web monitor application which I will be hosting in the cloud. For creating the device application, start with selection the Windows Universal project templates and then select Blank App (Universal) template.

There is a project template available for Visual Studio which will let you build applications for Raspberry PI. The problem with that project type is that you will not be able to make use of the libraries or the SDK available for .NET yet. You can still build applications with this project type but you have to make use of REST API and using REST API is not as straight-forward as SDKs.

After the project creation is successful, you need to add below dependencies which will let you connect to IoT Hub and have a device to cloud communication. Below are the NuGet packages you need to install.

This package will enable you to have a device to cloud communication.

Install-Package BuildAzure.IoT.Adafruit.BME280

This package is the wrapper for the BME280 sensor.

Install-Package Microsoft.Azure.Devices.Client

Add the below namespaces in the Main.xaml.cs file.

using Newtonsoft.Json;
using Microsoft.Azure.Devices.Client;
using BuildAzure.IoT.Adafruit.BME280;

Declare the variables.

private DeviceClient _deviceClient;
private BME280Sensor _sensor;
private string iotHubUri = "piiothub.azure-devices.net";
private string deviceKey = "Tj60asOk5ffVAT6a6SvZKMOqo8DYKSwWV7eQ2pLf0/k=";

The iothuburi is the URL for the IoT Hub created in the Azure portal or from the IoT dashboard. You can get the device key from the Device Explorer section in the IoT Hub. You can see how to get the Key from the Device Explorer from the below screen shot. I am using the Primary Key as a devicekey in my code.

In the constructor, initialize the DeviceClient and BME280Sensor objects. I have commented out the InitializeComponent(), but if you want you can keep it as it is.

public MainPage()
{
    //this.InitializeComponent();
    deviceClient = DeviceClient.Create(iotHubUri, new DeviceAuthenticationWithRegistrySymmetricKey("raspberrypi", deviceKey), TransportType.Mqtt);
    _sensor = new BME280Sensor();
    DeviceToCloudMessage();
}

Now add a function named DeviceToCloudMessage which will connect to the IoT Hub using the device key, read the sensor data, serialize the data in JSON format and send it to the IoT Hub. Add this at the very bottom of the constructor.

private async void DeviceToCloudMessage()
{
    await _sensor.Initialize();
    float temperature = 0.00f;
    float humidity = 0.00f;
    while (true)
    {
        temperature = await _sensor.ReadTemperature();
        humidity = await _sensor.ReadHumidity();
        var sensorData = new
        {
            date = String.Format("{0}, {1}, {2}",
                                 DateTime.Now.ToLocalTime().TimeOfDay.Hours,
                                 DateTime.Now.ToLocalTime().TimeOfDay.Minutes,
                                 DateTime.Now.ToLocalTime().TimeOfDay.Seconds),
            temp = Math.Round(temperature, 2),
            humid = Math.Round(humidity, 2)
        };
        var messageString = JsonConvert.SerializeObject(sensorData);
        var message = new Message(Encoding.ASCII.GetBytes(messageString));
        await deviceClient.SendEventAsync(message);
        //Debug.WriteLine("{0} > Sending message: {1}", DateTime.Now, messageString);
        Task.Delay(5000).Wait();
    }
}

The date property in the sensorData type is being set in this specific way because we want to see the graph continuously moving. I can also read Pressure from the sensor but as I am not interested in showing this data, I am skipping it out. If you want you can use it but you also have to change the web app to show this reading. Before I can send this data to the IoT Hub, I am serializing the sensorData object and use ASCII encoding to get the byte[] and pass it to the Message class constructor. The last step is to send the message to the hub by using the DeviceClient class SendEventAsync method. In the last line I am adding a delay of 5 seconds between each reading. You can increase the time and see how the chart renders. I recommend not to go below this time delay as this might give you some false readings from the sensor.

I took this code from the documentation here and tweaked it to have the real data from the sensor connected to the device. Press Ctrl+Shift+B to build the solution. Navigate to Project and then click <project> Properties. Under Debug section, set the Start Options as shown below. Notice that the Platform is set to ARM and Configuration is Active (Debug). When you are done with debugging the code and ready to deploy the code to the device you need to change it to Release.

Unlike Raspbian, where we ssh or RDP into the device and compile or build the code, Windows 10 IoT works in a different way. You need to remote deploy the application from Visual Studio. Also, if you have enabled the remote debugging on the device, you also have to add the port number to the Remote Machine value and set the Authentication Mode to None.

NOTE: When you make changes to device application, you also need to change/increment the version of the application in the Package.appxmanifest file. If you miss this step then the application deployment will fail.

To deploy the app to the device, click the Build menu and select Deploy <Project Name>. If you are deploying the application for the first time, it will take some time to get deployed. After the deployment is successful, you will see the application under the App section. At this moment, the application is stopped. To start the application in the, click the Play button in the list to start the application. The App Type show the type of application or you can say the mode it is running on. When I created the application, I selected the UWP application and therefore, it will always run in the Foreground type. The reason I choose this application type (UWP) is because I can then use the SDK to communicate with the IoT Hub. If you want to create a Background application type then you can download this Visual Studio Extension for 2017 and this for Visual Studio 2015 which will let you build the background application. Keep in mind that you cannot use any SDK library with this project type. All you have now is the power of REST API which is not very easy to use.

Setting the Web App

You can build any kind of app to visualize the data, but because the SDK support for .NET is good I am going to set up a simple MVC application using .NET Framework. As of now, you cannot use Azure SDK with .NET Core application and also it looks like the team does not have any plans to shift their focus for releasing the SDK for .NET Core anytime soon.

Start with installing NuGet packages. First with the package that will let you read the data from the Queue.

Install-Package WindowsAzure.ServiceBus

Because this is a real-time application you also need to install SignalR

Install-Package Microsoft.AspNet.SignalR

After these packages are installed, You need a way to visualize the temperature and humidity received from the queue. To do this you can use any jquery chart plugin or any other library of your choice. I will be using Google Charts in my app, Line Charts to be precise. Installation is straight forward for the charts and you can play around with different options to tweak the look and feel of the chart. I am going to add the chart code in the Index.cshtml file. Below is the complete code for the chart to render.

<script type="text/javascript" src="https://www.gstatic.com/charts/loader.js"></script>
<div id="chart_div"></div>
<script>
    var data = [];
    var chart;
    google.charts.load('current', { packages: ['corechart', 'line'] });
    google.charts.setOnLoadCallback(loadChart);
    var hub = $.connection.ioTHub;
    $.connection.hub.start();
    hub.client.iotHubNotification = function (d) {
        console.log(d);
        var pi = JSON.parse(d);
        var time = pi.date.split(',');
        var temp = pi.temp;
        var humid = pi.humid;
        data.addRows([[[parseInt(time[0]), parseInt(time[1]), parseInt(time[2])], temp, humid]]);
        var options = {
            height: 250,
            hAxis: {
                title: 'Time'
            },
            vAxis: {
                title: 'Temperature / Humidity',
                gridlines: { count: 22 }
            }
        };
        chart.draw(data, options);
    };
    function loadChart() {
        data = new google.visualization.DataTable();
        data.addColumn('timeofday', 'Time');
        data.addColumn('number', 'Temperature');
        data.addColumn('number', 'Humidity');
        var options = {
            height: 250,
            hAxis: {
                title: 'Time'
            },
            vAxis: {
                title: 'Temperarture / Humidity',
                gridlines: { count: 22 }
            }
        };
        chart = new google.visualization.LineChart(document.getElementById('chart_div'));
        chart.draw(data, options);
    }
</script>

I make use of the sample code from the examples at Google Charts and used it along with some minor tweaks to suite my needs. loadChart() function is called for the first time when the page is loaded and then I have a SignalR hub which updates the chart with the same option sets that I have in the loadChart() function. I add 3 data columns to display the data in the chart. There are few things that I would like to talk about in the above code. First is the timeofday is displayed on a X-Axis because time is continuous and I want to update the chart with time. Second, the gridlines property of the vAxis let you set how many horizontal rows you want to see in the graph. I have increased it to quite a significant number because I want to see the graph in more detail. You can play around with these settings and see what looks better for you. Third, the response I receive in the hub is in string format and therefore, I have parsed that string to JSON in order to read it and set to the rows. Also note that I have leave the console.log in the above code so that when you run the web app you can see the raw data in the console.

In the HomeController.cs, first I will set the connection string to the queue which I have associated with the IoT Hub endpoint in Azure and set the queue name along with the IHubContext for SignalR communication.

private string connectionString = "Endpoint=sb://iothubqueue-ns.servicebus.windows.net/;SharedAccessKeyName=RootManageSharedAccessKey;SharedAccessKey=88kJcD1mvJnO1jtiiY+AcUtIoinW//V/lF2WicOJ50s=";
private string queueName = "iothubqueue";
private IHubContext _hubContext;

In the constructor, initialize the IHubContext object.

_hubContext = GlobalHost.ConnectionManager.GetHubContext<IoTHub>();

Here is the code for the IoTHub class:

using Microsoft.AspNet.SignalR;
namespace IoTHubTempWebApp.Hubs
{
    public class IoTHub : Hub
    {
        public void IoTHubNotification(string value)
        {
            Clients.All.iotHubNotification(value);
        }
    }
}

And the code for Startup class (Startup.cs)

using Microsoft.Owin;
using Owin;
[assembly: OwinStartup(typeof(IoTHubTempWebApp.Startup))]
namespace IoTHubTempWebApp
{
    public class Startup
    {
        public void Configuration(IAppBuilder app)
        {
            app.MapSignalR();
        }
    }
}

In the Index ActionResult, I will start a task which will read the messages from the queue and then SignalR will broadcast it to the client side which in turn update the chart with the latest data or temperature/humidity readings.

public ActionResult Index()
{
    Task task = Task.Run(() =>
    {
        QueueClient client = QueueClient.CreateFromConnectionString(connectionString, queueName, ReceiveMode.ReceiveAndDelete);
        client.OnMessage(message =>
        {
            Stream stream = message.GetBody<Stream>();
            StreamReader reader = new StreamReader(stream, Encoding.ASCII);
            string s = reader.ReadToEnd();
            _hubContext.Clients.All.ioTHubNotification(s);
        });
    });
    task.Wait();
    return View();
}

First I create a QueueClient object with the connection string and queue name along with the ReceiveMode. I have set ReceiveMode to ReceiveAndDelete as I want to delete the data from the queue once it has been read. Though you have an option to have the data in the queue for the maximum of 7 days in Azure. The QueueClient has an OnMessage event which process a message in an event-driven message pump, which means that as soon as the something is being added to the queue, this event is fired. I get the message body in the form of Stream, read the message to the end and pass the message to SignalR hub which in turn updates my chart in real-time. Here is the final output I have now.

If you look closely, you will notice that the temperature line is not as smooth as you thought it would be. This is because every time when Raspberry PI read the data from the sensor there is a slight change in the temperature only with a few decimal places. If you don't like this then you can change the gridlines property of the vAxis in the chart. Also you can try changing the time delay on the device to see if that impacts the lines on the chart. try changing both time delay in the device application and the gridlines in the web application to see how the chart renders.

The complete source code is available on Github.

 No Rating

Upload Files In .NET Core By Drag And Drop Using Dropzone.JS

17. October 2016 21:41

.NET Core ASP.NET MVC Jquery Web 

Mostly all web applications out there has some way or the other amazing ways to upload a single or multiple files. While surfing on Github I found this amazing library to upload the files to the server in a unique way with lot of configurations. It support parallel uploads along with cancellation of the files which are in the upload queue along with a good looking progress bar to show the progress of upload.

I get the drag and drop to work in just like 5 minutes. It is super easy and that with some powerful configurations. To install Dropzone you can use the Nuget command.

PM> Install-Package dropzone
Add reference of js and css files on your page. To get the UI ready use this HTML.
<div class="row">
    <div class="col-md-9">
        <div id="dropzone">
            <form action="/Home/Upload" class="dropzone needsclick dz-clickable" id="uploader">
                <div class="dz-message needsclick">
                    Drop files here or click to upload.<br>
                </div>
            </form>
        </div>
    </div>
</div>

The output of the above HTML looks like this.

There are few points to be noted here in the above HTML. Notice the action and class attributes for the form element. You will also be needing the id attribute as well. Here the action attribute points to the ActionResult which is responsible to handle the file upload. I have pointed it to the Upload ActionResult in my controller class which accepts a parameter of type IFormFile. In case of a MVC application, we would have used HttpPostedFileBase class. Here is the complete code which handles the file upload.

[HttpPost]
public async Task<IActionResult> Upload(IFormFile file)
{
    var uploads = Path.Combine(_environment.WebRootPath, "uploads");
    if (file.Length > 0)
    {
        using (var fileStream = new FileStream(Path.Combine(uploads, file.FileName), FileMode.Create))
        {
            await file.CopyToAsync(fileStream);
        }
    }
    return RedirectToAction("Index");
}

The _environment variable you see in the above code is the instance of IHostingEnvironment interface which I have injected in the controller. Always use IHostingEnvironment interface to resolve the paths. You can hard code the file path but it may not work on other platforms. The combine method returns the correct path based on the platform the code is executing. The WebRootPath property returns the path of wwwroot folder and the second parameter uploads is then appended correctly with the path.

Now it is time to make some adjustment in the Dropzone configuration. Recall the id attribute of the form element. I named it uploader. The Dropzone library has a unique way to set the configuration. Like this.

<script>
    $(document).ready(function () {
        Dropzone.options.uploader = {
            paramName: "file",
            maxFilesize: 2,
            accept: function (file, done) {
                if (file.name == "test.jpg") {
                    alert("Can't upload a test file.");
                }
                else {
                    //Show a confirm alert and display the image on the page.
              }
           }
        };
});
</script>
You have to be a bit careful when setting this configuration. In the configuration above the paramName states the name that will be used to transfer the file. This name should be the same as the IFormFile parameter of the Upload method in the controller. In this case I am using file and the same has to be there in the param of the Upload method. If the names mis-match the files will not be uploaded. The other parameter I am using is the maxFileSize and is very much self-explanatory. I have set the size to be 2 MB and because of this configuration, any file above this limit will not be uploaded. Here is an example of such kind of failure.
All the other files were uploaded successfully except one file which is 4.32 MB and is way beyond the limit I set in my Dropzone configuration. If you hover the file, you will see why it got failed.
This is the simplest approach through which you can have drag and drop upload support in your applications. The configuration I am using here is the simplest and minimalistic configuration that can get you started in no time. There are some more powerful configurations like parallelUploads and uploadMultiple that you should look into and use it.
 No Rating

Get The Correct File Paths In .NET Core

3. October 2016 22:12

.NET Core ASP.NET MVC Ubuntu Web 

I run a custom blog engine I wrote myself on MVC 4 and as you would have thought it is on Windows hosting. I am re-writing my blog engine in .NET Core so that I can get it running on a Linux hosting as well. I am using Windows machine and Visual Studio Community Edition to write it and using a Ubuntu VM box to test it. My current blog uses SQL Server as a back-end but Linux does not support SQL Server and therefore I have to move to MySQL or any other No-SQL database available for Linux distros. For now I will be saving everything on the disk in JSON format.

The problem comes when I switch from database to file system. Both Windows and Linux file systems are different, so when working I hard-coded the data folder which holds all the posts for my blog. I publish the solution and deploy it to Ubuntu VM. I started the server and it ended up showing me this error in the console.

Notice the path of the data folder in the above screenshot. The path I am referring in the code ends up with backward slash which is not a UNIX format. Also a thing to note here is that UNIX follows a strict naming convention unlike Windows. For example, in Windows Data and data are same, but in UNIX they are not. Because I have hard-coded the path in my application while developing, it will end up in an error on UNIX machine.

To make the paths consistent in your application, I have to make use of the System.IO namespace Path class's Combine method. Combine method is an overloaded method. If you are hearing this method for the first time then make sure you read the documentation. This method renders the correct path based on the platform my code is executing. So instead of doing something like this.

var _posts = Directory.EnumerateFiles("Data\\Posts");
I have done something like this.
string posts = Path.Combine("Data", "Posts");
_files = Directory.EnumerateFiles(posts);
In short, if you are developing a .NET Core application targeting Windows along with UNIX and OSX, then make sure that you follow this approach to get the correct paths or else you will end up with an exception and your website inaccessible.
 No Rating

Building Angular 2 App With Web API And .NET Core

21. May 2016 18:45

.NET Core ASP.NET MVC 

Setting up the new Angular 2 app with Web API and .NET core is easy but can be a bit tricky. The older beta releases of Angular 2 works fine as there are not many files to refer and to work with. When I started using Angular 2 it was in RC1 and the way the files are being referenced in the app is bit different than the older versions of Angular 2. I don't want to repeat these steps again and again so I put up a Github repo for this seed project. You clone it and hit F5 and you will have your Angular 2 app with Web API.

At the time of writing this post I am using .NET Core version 1.0.0-preview1-002702. The complete seed project is available on GitHub.
Here is how I did it. Select ASP.NET Core Web Application (.NET Core). I have installed the new .NET Core RC2. Name the project as you like.
Select the Web API project template.
After the project creation is successfull. The first thing is to create a Views folder. The folder structure is like the same like it is for MVC application. This is how the folder structure looks like.
Because its a a Web API project, by default it will not render the views and therefore we have to add some dependencies in project.json file.
"Microsoft.AspNetCore.StaticFiles": "1.0.0-rc2-final",
"Microsoft.AspNetCore.Mvc.TagHelpers": "1.0.0-rc2-final",
"Microsoft.AspNetCore.Mvc.WebApiCompatShim": "1.0.0-rc2-final"
The project.json file is different than that of the older version of the .NET Core. You will see the difference when you see it.
Next we set the routes for the views we have in the Startup.cs file.
app.UseMvc(routes =>
{
    routes.MapRoute("default",
                    "{controller=Home}/{action=Index}/{id?}");

    routes.MapWebApiRoute("defaultApi",
                          "api/{controller}/{id?}");
});
The routes are now set and you can run the application and check if you can see the view or not. Once that is set, let start adding the support for Angular. Here is the list of the files which we need to add.
tsconfig.json file

Add the below code to the file and save it.
{
  "compilerOptions": {
    "target": "es5",
    "module": "system",
    "moduleResolution": "node",
    "sourceMap": true,
    "emitDecoratorMetadata": true,
    "experimentalDecorators": true,
    "removeComments": false,
    "noImplicitAny": false,
    "rootDir": "wwwroot",
    "outDir": "wwwroot",
    "listFiles": true,
    "noLib": false,
    "diagnostics": true
  },
  "exclude": [
    "node_modules"
  ]
}
typings.json file

Add a blank .json file and name it typings.json

Add below lines to the typings.json file.
{
  "ambientDependencies": {
    "es6-shim": "registry:dt/es6-shim#0.31.2+20160317120654",
    "jasmine": "registry:dt/jasmine#2.2.0+20160412134438"
  }
}
As per the official Angular documentation, we will stick with NPM to fulfill the client-side dependencies. Start with adding new npm Configuration File. The content of the NPM file is almost the same, but I made few changes as per my requirement. Here is the complete configuration file.
{
  "name": "Angular2WebAPI-Seed",
  "version": "1.0.0",
  "scripts": {
    "postinstall": "typings install"
  },
  "license": "ISC",
  "dependencies": {
    "@angular/common": "2.0.0-rc.1",
    "@angular/compiler": "2.0.0-rc.1",
    "@angular/core": "2.0.0-rc.1",
    "@angular/http": "2.0.0-rc.1",
    "@angular/platform-browser": "2.0.0-rc.1",
    "@angular/platform-browser-dynamic": "2.0.0-rc.1",
    "@angular/router": "2.0.0-rc.1",
    "@angular/router-deprecated": "2.0.0-rc.1",
    "@angular/upgrade": "2.0.0-rc.1",

    "lodash": "4.12.0",
    "systemjs": "0.19.27",
    "es6-shim": "^0.35.0",
    "reflect-metadata": "^0.1.3",
    "rxjs": "5.0.0-beta.6",
    "zone.js": "^0.6.12",
    "angular2-in-memory-web-api": "0.0.7",
    "bootstrap": "^3.3.6"
  },
  "devDependencies": {
    "gulp": "3.9.1",
    "concurrently": "^2.0.0",
    "typescript": "^1.8.10",
    "typings": "^0.8.1"
  }
}
Notice the postinstall section in the npm configuration file. In this section we are installing the types required by our Angular 2 application. Now we can start setting up the Angular stuff in the wwwroot folder. Create an app and js folder inside this folder. Inside app folder create a new .ts file (TypeScript file). Here is the folder structure looks like in the wwwroot folder.
You can see a system.config.js file is the same as you can see the Angular quickstart guide. I just have map the paths for the dependencies so that can be loaded without any problem. The main.ts file will be the main bootstrapper and app.component.ts file is the component file which will render the content on the page. At this point running the application will fail and it will give you several warnings and errors. To resolve that we need to add the references and we can do this easily by using a gulp file. The gulp file will automate the copying of dependencies in the wwwroot folder and ease our task. 
The content of the gulp file in this case looks like this.
/// <binding BeforeBuild='default' />

var _ = require('lodash');
var gulp = require('gulp');

var js = [
    'node_modules/zone.js/dist/zone.min.js',
    'node_modules/systemjs/dist/system.js',
    'node_modules/reflect-metadata/Reflect.js',
    'node_modules/es6-shim/es6-shim.min.js'
];

var map = [
    'node_modules/es6-shim/es6-shim.map',
    'node_modules/reflect-metadata/reflect.js.map',
    'node_modules/systemjs/dist/system.js.map'
];

var folders = [
    'node_modules/@angular/**/*.*',
    'node_modules/rxjs/**/*.*'
];

gulp.task('copy-js', function () {
    _.forEach(js, function (file, _) {
        gulp.src(file)
       .pipe(gulp.dest('./wwwroot/js'))
    });
});

gulp.task('copy-map', function () {
    _.forEach(map, function (file, _) {
        gulp.src(file)
        .pipe(gulp.dest('./wwwroot/js'))
    });
});

gulp.task('copy-folders', function () {
    _.forEach(folders, function (folder) {
        gulp.src(folder, { base: 'node_modules' })
            .pipe(gulp.dest('./wwwroot/js'))
    });
})

gulp.task('default', ['copy-js', 'copy-map', 'copy-folders']);
Notice the very first line in the file above. We want to execute the task on every time before the build is triggered. The gulp file will only be copying the required files to the js folder and other unnecessary files will not be required. After the task is executed here is the how the final directiry structure will look like under wwwroot.
Depending on the selector you have in your component, you need to add the selector to your view. In my case the selector is app, hence add <app></app> in your Index.cshtml file.

After you set up the index page, there is one more thing that you have to do is to set the launch URL.

The URL may change and differ as per your API endpoint. Press F5 to run the application.

Currently rated 5.0 by 8 people

Parsing Markdown Using Custom TagHelper In ASP.NET MVC 6

30. November 2015 21:01

.NET Core ASP.NET MVC 

Previous versions of MVC allows us to write HtmlHelpers which does a pretty good job then and they are doing it now as well. But in MVC 6, the ASP.NET team has introduced TagHelpers.

Parsing Mardown in .NET is way too simple than one can imagine. Thanks to Stackoverflow's MardownSharp and Karlis Gangis's CommonMark.NET. I use CommonMark.NET as it provides a much faster parsing than other libraries. The blogging platform I use is a custom blogging engine I wrote in MVC4. The post content is saved in HTML which makes my raw HTML way to messy when compared to simple markdown syntax. I have no plans to change the way it is right now, but for the other simple application which is quite similar to notes taking or blogging apps, I would like to save the content in markdown.

I will start with a simple implementation of this custom TagHelper and then then we can look into the other ways to enhance it. Here is how easy it is to create your custom TagHelper.

Create a new class file MarkdownTagHelper.cs. Inside the file, rename the class name to match the file name or you can change the name way you like. In my case I am keeping the class name as same as the file name.

Pay attention to the name of the custom TagHelper. By design, compiler will remove the word TagHelper from the class name and the rest of the name will become your custom TagHelper name.

The next step is to inherit our class with TagHelpers class. Every custom TagHelper will inherit this class. Just like the UserControl class when creating a custom user control. The TagHelper provide us two virtual methods, Process and a ProcessAsync method which we will be going to override to implement our custom logic for our custom markdown TagHelper. The first parameter is the TagHelperContext which holds the information about the current tag and the second parameter is TagHelperOutput object represents the output being generated by the TagHelper. As we need to parse the markdown in our razor view pages, we need to add reference of CommonMark.Net library. Use the below Nuget command to add it to your current project.

Install-Package CommonMark.Net

Till here this is how the code will look like.

So now we have our custom TagHelper that will let us parse the markdown. But to use it in our views we need to opt-in for this TagHelper in the _ViewImports.cshtml file. To enable your custom TagHelper just type in like this:

@addTagHelper "*, WebApplication1"

Your custom tag helper would have been turned purple in color on the view page. It is similar to the line above it where @addTagHelper is importing all the TagHelpers from the given namespace. If you are not interested in opting-in for all the TagHelpers in the given namespace then make use of the @removeTagHelper to disable the TagHelpers you don’t need. For this I want all the tag helpers I have created to be a part of the application and hence the * symbol.

In your view, where you want to use this just type in <markdown> and inside this tag you should have your markdown. To test it, you can view the any raw file in Github and copy the text. I am using README.md from CommonMark.NET and it rendered perfectly.

Caution: When copy-pasting the markdown code from anywhere to your view make sure that you do not have a whitespace in the front of the line. This is only applicable when you are working with the in-line markdown. Here is the screenshot with comparison.

Hit F5 and see the markdown tag helper in action. Below is the output I get.

This is the simplest of all. Now let’s add some prefix to our custom TagHelper. To add a custom tag prefix to the TagHelper, we just need to pay a visit to _ViewImports.cshtml file again and add a new line like so:

@tagHelperPrefix "blog-"

After adding the above line in the file, go to the view page where you have used your custom TagHelper and you can notice that the <markdown> tag isn’t purple anymore. This is because we now have to add a custom tag-prefix that we just defined in the _ViewImports.cshtml file. Change it from <markdown> to <blog-markdown> and it is purple again.

By design, the TagHelper will take <markdown> as a tag to be processed. But it can be easily ignored by using the HtmlTargetElement attribute at the top of the class and allow the use of another name rather than <markdown>. This does not mean that you cannot use <markdown> but instead you can also use the custom TagHelper with the name specify in this attribute.

Now let’s add some attributes to my markdown TagHelper. Let’s try to add a url attribute which will help user to render the markdown on the view from a remote site like Github. To add an attribute, simply add a new public property of type string and call it url. When you create a public property in the TagHelper class it is automatically assumed it is an attribute. To make use of this property, my view now simply say this:

<blog-markdown url="https://raw.githubusercontent.com/Knagis/CommonMark.NET/master/README.md">
</blog-markdown>

The url attribute value is being read by the TagHelper which in turn read the whole string of markdown from Github and render the HTML on the page. Let’s focus again on TargetElement attribute for a while. Consider a scenario where you don’t want your custom TagHelper to render or work if the attributes are not passed or missing. This is where HtmlTargetElement attribute comes into picture. If I don’t want my TagHelper to work if the url parameter is missing then you can simple write your HtmlTargetElement attribute like so:

[HtmlTargetElement("markdown", Attributes = "url")]

Notice the Attributes parameter. The Attributes parameter allows you to set the name of all the attributes which should be processed by your TagHelper or else the TagHelper will not work. For instance, if I just use the <markdown> TagHelper but did not pass the url attribute, the TagHelper will not execute and you will see the raw markdown code. My requirement is to have this TagHelper working with or without the use of url attribute. I can comment out or remove the HtmlTargetElemen attribute or just remove the Attributes parameter to get going.

Here is the complete MarkdownTagHelper.cs:

using Microsoft.AspNet.Razor.Runtime.TagHelpers;
using System;
using System.Net.Http;
using System.Threading.Tasks;

namespace WebApplication1.TagHelpers
{
    //[HtmlTargetElement("markdown", Attributes = "url")]
    public class MarkdownTagHelper : TagHelper
    {
        //Attribute for our custom markdown
        public string Url { get; set; }

        private string parse_content = string.Empty;

        //Stolen from: http://stackoverflow.com/questions/7578857/how-to-check-whether-a-string-is-a-valid-http-url
        private bool isValidURL(string URL)
        {
            Uri uriResult;
            return Uri.TryCreate(URL, UriKind.Absolute, out uriResult)
                && (uriResult.Scheme.ToLowerInvariant() == "http" || uriResult.Scheme.ToLowerInvariant() == "https");
        }

        public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
        {
            if (context.AllAttributes["url"] != null)
            {
                string url = context.AllAttributes["url"].Value.ToString();
                string webContent = string.Empty;
                if (url.Trim().Length > 0)
                {
                    if (isValidURL(url))
                    {
                        using (HttpClient client = new HttpClient())
                        {
                            webContent = await client.GetStringAsync(new Uri(url));
                            parse_content = CommonMark.CommonMarkConverter.Convert(webContent);
                            output.Content.SetHtmlContent(parse_content);
                        }
                    }
                }
            }
            else
            {
                //Gets the content inside the markdown element
                var content = await output.GetChildContentAsync();

                //Read the content as a string and parse it.
                parse_content = CommonMark.CommonMarkConverter.Convert(content.GetContent());

                //Render the parsed markdown inside the tags.
                output.Content.SetHtmlContent(parse_content);
            }
        }
    }
}

I found the full TagHelper feature in MVC 6 a lot more convenient and powerful. I hope you like it as well.

 No Rating

«123456»