In this chapter, we will cover the following recipes:

• Building a backend Web API using HTTP triggers
• Persisting employee details using Azure Storage table output bindings
• Saving the profile images to Queues using Queue output bindings
• Storing the image in Azure Blob storage
• Cropping an image using ImageResizer trigger

Every software application needs backend components that are responsible for taking care of the business logic and storing the data into some kind of storage such as database, filesystem, and so on. Each of these backend components could be developed using different technologies. Azure serverless technology also allows us to develop these backend APIs using Azure Functions.

Azure Functions provide many out-of-the-box templates that solves most of the common problems such as connecting to storage, building Web APIs, cropping the images, and so on. In this chapter, we will learn how to use these built-in templates. Along with learning the concepts related to Azure serverless computing, we will also try to implement a solution to a basic domain problem of creating components required for any organization to manage the internal employee information.

Below is a simple diagram that helps you understand what we will be going to achieve in this chapter:

We will use Azure serverless architecture for building a Web API using HTTP triggers. These HTTP triggers could be consumed by any frontend application that is capable of making HTTP calls.

Let's start our journey of understanding Azure serverless computing using Azure Functions by creating a basic backend Web API that responds to HTTP requests:

• Please refer to the URL https://azure.microsoft.com/en-in/free/?&wt.mc_id=AID607363_SEM_8y6Q27AS for creating a free Azure Account.
• Also, visit https://docs.microsoft.com/en-us/azure/azure-functions/functions-create-function-app-portal to understand the step by step process of creating a function app and https://docs.microsoft.com/en-us/azure/azure-functions/functions-create-first-azure-function to create a function. While creating a function, a Storage Account is also created for storing all the files. Please remember the name of the Storage Account which will be used later in the other chapters.

1. Navigate to the Function App listing page. Choose the function app in which you would like to add a new function.
2. Create a new function by clicking on the + icon as shown in the following screenshot:

3. If you have created a brand new function, then clicking on the + icon in the preceding step, you would see the Get started quickly with a premade function page. Please click on the create your own custom functions link to navigate to the page where you can see all the built-in templates for creating your Azure Functions.
4. In the Choose a template below or go to the quickstart section, choose HTTPTrigger-CSharp as shown in the following screenshot to create a new HTTP trigger function:

5. Provide a meaningful name. For this example, I have used RegisterUser as the name of the Azure Function.
6. In the Authorization level drop-down, choose the Anonymous option as shown in the following screenshot. We will learn more about the all the authorization levels in Chapter 9, Implement Best Practices for Azure Functions:

7. Once you provide the name and choose the Authorization level, click on Create button to create the HTTP trigger function.
8. As soon as you create the function, all the required code and configuration files will be created automatically and the run.csx file will be opened for you to edit the code. Remove the default code and replace it with the following code:

        using System.Net;
        public static async Task<HttpResponseMessage> 
        Run(HttpRequestMessage req, TraceWriter log)
        {
            string firstname=null,lastname = null;
            dynamic data = await req.Content.ReadAsAsync<object>();
            firstname = firstname ?? data?.firstname;
            lastname = data?.lastname;
            return (lastname + firstname) == null ? 
             req.CreateResponse(HttpStatusCode.BadRequest, 
             "Please pass a name on the query string or in the 
             request body") : 
            req.CreateResponse(HttpStatusCode.OK, "Hello " + 
             firstname + " " + lastname);
        }

9. Save the changes by clicking on the Save button available just above the code editor.
10. Let's try to test the RegisterUser function using the Test console. Click on the tab named Test as shown in the following screenshot to open the Test console:

11. Enter the values for firstname and lastname, in the Request body section as shown in the following screenshot:

Please make sure you select POST in the HTTP method drop-down.

12. Once you have reviewed the input parameters, click on the Run button available at the bottom of the Test console as shown in the following screenshot:

13. If the input request workload is passed correctly with all the required parameters, you will see a Status 200 OK, and the output in the Output window will be as shown in the preceding screenshot.

We have created the first basic Azure Function using HTTP triggers and made a few modifications to the default code. The code just accepts firstname and lastname parameters and prints the name of the end user with a Hello {firstname} {lastname} message as a response. We have also learnt how to test the HTTP trigger function right from the Azure Management portal.

• The Enabling authorization for function apps recipe in Chapter 9, Implement Best Practices for Azure Functions

In the previous recipe, you have learnt how to create an HTTP trigger and accept the input parameters. Let's now work on something interesting, that is, where you store the input data into a persistent medium. Azure Functions supports us to store data in many ways. For this example, we will store the data in Azure Table storage.

In this recipe, you will learn how easy it is to integrate an HTTP trigger and the Azure Table storage service using output bindings. The Azure HTTP trigger function receives the data from multiple sources and stores the user profile data in a storage table named tblUserProfile.

• For this recipe, we will use the same HTTP trigger that we have created in our previous recipe.
• We will be using Azure Storage Explorer which is a tool that helps us to work with the data stored in Azure Storage account. You can download it from http://storageexplorer.com/.
• You can learn more about Connect to the Storage Account using Azure Storage Explorer at https://docs.microsoft.com/en-us/azure/vs-azure-tools-storage-manage-with-storage-explorer

1. Navigate to the Integrate tab of the RegisterUser HTTP trigger function.
2. Click on the New Output button and select Azure Table Storage then click on the Select button:

3. Once you click on the Select button in the previous step, you will be prompted to choose the following settings of the Azure Table storage output bindings:
   • Table parameter name: This is the name of the parameter that you will be using in the Run method of the Azure Function. For this example, please provide objUserProfileTable as the value.
   • Table name: A new table in the Azure Table storage will be created to persist the data. If the table doesn't exist already, Azure will automatically create one for you! For this example, please provide tblUserProfile as the table name.
   • Storage account connection: If you don't see the Storage account connection string, click on the new (shown in the following screenshot) to create a new one or to choose an existing storage account.
   • The Azure Table storage output bindings should be as shown in the following screenshot:

4. Click on Save to save the changes.

5. Navigate to the code editor by clicking on the function name and paste the following code:

        #r "Microsoft.WindowsAzure.Storage"
        using System.Net;
        using Microsoft.WindowsAzure.Storage.Table;

        public static async Task<HttpResponseMessage>
        Run(HttpRequestMessage req,TraceWriter
        log,CloudTable objUserProfileTable)
        {
            dynamic data = await 
             req.Content.ReadAsAsync<object>();
            string firstname= data.firstname;
            string lastname=data.lastname;
 
            UserProfile objUserProfile = new UserProfile(firstname,    
             lastname);
            TableOperation objTblOperationInsert =  
             TableOperation.Insert(objUserProfile);
            objUserProfileTable.Execute(objTblOperationInsert);
    
            return req.CreateResponse(HttpStatusCode.OK,
             "Thank you for Registering..");
        } 

        public class UserProfile : TableEntity
        {
            public UserProfile(string lastName, string firstName)
            {
                this.PartitionKey = "p1";
                this.RowKey = Guid.NewGuid().ToString();;
                this.FirstName = firstName;
                this.LastName = lastName;
            }
            public UserProfile() { }
            public string FirstName { get; set; }
            public string LastName { get; set; }
        }

6. Let's execute the function by clicking on the Run button of the Test tab by passing firstname and lastname parameters in the Request body as shown in the following screenshot:

7. If everything went well, you should get a Status 200 OK message in the Output box as shown in the preceding screenshot. Let's navigate to Azure Storage Explorer and view the table storage to see if the table named tblUserProfile was created successfully:

Azure Functions allows us to easily integrate with other Azure services just by adding an output binding to the trigger. For this example, we have integrated the HTTP trigger with the Azure Storage table binding and also configured the Azure Storage account by providing the storage connection string and the Azure Storage table name in which we would like to create a record for each of the HTTP requests received by the HTTP trigger.

We have also added an additional parameter for handling the table storage named objUserProfileTable, of type CloudTable, to the Run method. We can perform all the operations on the Azure Table storage using objUserProfileTable.

We have also created an object of UserProfile, and filled it with the values received in the request object, and then passed it to a table operation. You can learn more about handling operations on Azure Table storage service from the URL https://docs.microsoft.com/en-us/azure/storage/storage-dotnet-how-to-use-tables.

When you create a new storage connection (please refer to the third step of the How to do it... section of this recipe) a new App settings will be created as shown in the following screenshot:

You can navigate to the App settings by clicking on Application settings of the Platform features tab as shown in the following screenshot:

Azure Table storage service is a NoSQL key-value persistent medium for storing semi-structured data. You can learn more about the same at https://azure.microsoft.com/en-in/services/storage/tables/.

The primary key of Azure Table storage tables has two parts as follows:

• Partition key: Azure Table storage records are classified and organized into partitions. Each record located in a partition will have the same partition key (p1 in our example).
• Row key: A unique value should be assigned for each of the rows.

Following is the very first line of the code in this recipe:

#r "Microsoft.WindowsAzure.Storage"

The preceding line of code instructs the function runtime to include a reference to the specified library to the current context.

In the previous recipe, you have learnt how to receive two string parameters firstname and lastname in the Request body, and store them in the Azure Table storage. In this recipe, you will learn how to receive a URL of an image and save the same in the Blob container of an Azure Storage account.

We could have processed the downloaded user profile image in the recipe Persisting employee details using Azure Storage table output bindings. However, keeping in mind the size of the profile pictures, the processing of images on the fly in the HTTP requests might hinder the performance of the function. For that reason, we will just grab the URL of the profile picture and store it in Queue, and later we can process the image and store it in the Blob.

We will be updating the code of the RegisterUser function that we have used in the previous recipes.

1. Navigate to the Integrate tab of the RegisterUser HTTP trigger function.
2. Click on the New Output button and select Azure Queue Storage then click on the Select button.
3. Provide the following parameters in the Azure Queue Storage output settings:
   • Queue name: Set the value of the Queue name as userprofileimagesqueue
   • Storage account connection: Please make sure that you select the right storage account in the Storage account connection field
   • Message parameter name: Set the name of the parameter to objUserProfileQueueItem which will be used in the Run method
4. Click on Save to the create the new output binding.

5. In this recipe, we will look at another approach of grabbing the request parameters for which we will use the Newtonsoft.JSON library to parse the JSON data. Let's navigate to the View files tab as shown in the following screenshot:

6. As shown in the preceding screenshot, click on Add to add a new file. Please make sure that you name it as project.json as shown in the preceding screenshot.
7. Once the file is created, add the following code to the project.json file. The following code adds the reference of the Newtonsoft.Json library.

        {
        "frameworks" : {
            "net46": {
                "dependencies":{
                    "Newtonsoft.Json" : "10.0.2"
                    }
                 }
            }
        }

8. Navigate back to the code editor by clicking on the function name (RegisterUser in this example) and paste the following code:

        #r "Microsoft.WindowsAzure.Storage"
        using System.Net;
        using Microsoft.WindowsAzure.Storage.Table;
        using Newtonsoft.Json;
        public static void Run(HttpRequestMessage req, 
                               TraceWriter log,
                               CloudTable         
                               objUserProfileTable,
                               out string 
                               objUserProfileQueueItem
                               )
        {
            var inputs = req.Content.ReadAsStringAsync().Result;
            dynamic inputJson = JsonConvert.DeserializeObject<dynamic>
             (inputs);
    
            string firstname= inputJson.firstname;
            string lastname=inputJson.lastname;
            string profilePicUrl = inputJson.ProfilePicUrl;

            objUserProfileQueueItem = profilePicUrl;
            UserProfile objUserProfile = new UserProfile(firstname, 
             lastname, profilePicUrl);
            TableOperation objTblOperationInsert = 
             TableOperation.Insert(objUserProfile);
           
            objUserProfileTable.Execute(objTblOperationInsert);
        }

        public class UserProfile : TableEntity
        {
            public UserProfile(string lastname, string firstname,
             string profilePicUrl)
            {
                this.PartitionKey = "p1";
                this.RowKey = Guid.NewGuid().ToString();
                this.FirstName = firstname;
                this.LastName = lastname;
                this.ProfilePicUrl = profilePicUrl;
            }
            public UserProfile() { }
            public string FirstName { get; set; }
            public string LastName { get; set; }
            public string ProfilePicUrl {get; set;}
        }

9. Click on Save to save the code changes in the code editor of the run.csx file.

10. Let's test the code by adding another parameter ProfilePicUrl to the Request body shown as follows then click on the Run button in the Test tab of the Azure Function code editor window: The image used in the below JSON might not exist when you are reading this book. So, Please make sure that you provide a valid URL of the image.

        {
            "firstname": "Bill",
            "lastname": "Gates",              
         "ProfilePicUrl":"https://upload.wikimedia.org/wikipedia/
          commons/1/19/Bill_Gates_June_2015.jpg"
        }

11. If everything goes fine you will see the Status : 200 OK message, then the image URL that you have passed as an input parameter in the Request body will be created as a Queue message in the Azure Storage Queue service. Let's navigate to Azure Storage Explorer, and view the Queue named userprofileimagesqueue, which is the Queue name that we have provided in the Step 3. Following is the screenshot of the Queue message that was created:

In this recipe, we have added Queue message output binding and made the following changes to the code:

• Added a reference to the Newtonsoft.Json NuGet library in the project.json file
• Added a new parameter named out string objUserProfileQueueItem which is used to bind the URL of the profile picture as a Queue message content
• We have also made the Run method synchronous by removing async as it doesn't allow us to have out parameters

The project.json file contains all the references of the external libraries that we may use in the Azure Function.

• The Persisting employee details using Azure Storage table Output Bindings recipe

Let's learn how to invoke an Azure Function when a new queue item is added to the Azure Storage Queue service. Each message in the Queue is the URL of the profile picture of a user which will be processed by the Azure Functions and will be stored as a Blob in the Azure Storage Blob service.

In the previous recipe, we have learnt how to create Queue output bindings. In this recipe, you will grab the URL from the Queue, create a byte array, and then write it to a Blob.

This recipe is a continuation of the previous recipes. Please make sure that you implement them.

1. Create a new Azure Function by choosing the QueueTrigger-C# from the templates.
2. Provide the following details after choosing the template:
   • Name your function: Please provide a meaningful name such as CreateProfilePictures.
   • Queue name: Name of the Queue which should be monitored by the Azure Function. Our previous recipe created a new item for each of the valid requests coming to the HTTP trigger (named RegisterUser) into the userprofileimagesqueue Queue. For each new entry of a queue message to this Queue storage, the CreateProfilePictures trigger will be executed automatically.
   • Storage account connection: Connection of the storage account where the Queues are located.
3. Review all the details, and click on Create to create the new function.
4. Navigate to Integrate tab then click on New Output then choose Azure Blob Storage then click on the Select button.
5. In the Azure Blob Storage output section, provide the following:
   • Blob parameter name: Set it to outputBlob
   • Path: Set it to userprofileimagecontainer/{rand-guid}
   • Storage account connection: Choose the storage account where you would like to save the Blobs:

6. Once you provide all the preceding details, click on the Save button to save all the changes.
7. Replace the default code of the run.csx file with the following code:

        using System;
        public static void Run(Stream outputBlob,string myQueueItem, 
         TraceWriter log)
        {
            byte[] imageData = null;
            using (var wc = new System.Net.WebClient())
            {
                imageData = wc.DownloadData(myQueueItem);
            }
        outputBlob.WriteAsync(imageData,0,imageData.Length);
        }

8. Click on the Save button to save the changes.
9. Let's go back to the RegisterUser function and test it by providing firstname, lastname, and ProfilePicUrl fields as we did in the Saving the profile images to Queues using Queue output bindings recipe.
10. Now, navigate to the Azure Storage Explorer, and look at the Blob container userprofileimagecontainer. You will find a new Blob as shown in the following screenshot:

11. You can view the image in any tool (such as MS Paint or Internet Explorer).

We have created a Queue trigger that gets executed as and when a new message arrives in the Queue. Once it finds a new Queue message, then it reads the message, and as we know the message is a URL of a profile picture. The function makes a web client request and downloads the image data in the form of byte array, and then writes the data into the Blob which is configured as an output Blob

The parameter rand-guid, will generate a new GUID and is assigned to the Blob that gets created each time the trigger is fired.

• The Building a backend Web API using HTTP triggers recipe
• The Persisting employee details using Azure Storage table output bindings recipe
• The Saving the profile images to Queues using Queue output bindings recipe
• The Storing the image in Azure Blob storage recipe

In the recent times, with the evolution of smart phones with high-end cameras, it's easy to capture a high-quality picture of huge sizes. It's good to have good quality pictures to refresh our memories. However, as an application developer or administrator, it would be a pain to manage the storage when your website is popular and you expect most of the users to get registered with a high-quality profile picture. So, it makes sense to use some libraries that could reduce the size of the high-quality images and crop them without losing the aspect ratio so that the quality of the image doesn't get reduced.

In this recipe, we will learn how to implement the functionality of cropping the image and reducing the size without losing the quality using one of the built-in Azure Function templates named ImageResizer .

In this recipe, you will learn how to use a library named ImageResizer. We will be using the library for resizing the image with the required dimensions. For the sake of simplicity, we will crop the image to the following sizes:

• Medium with 200*200 pixels
• Small with 100*100 pixels

1. Create a new Azure Function by choosing the Samples in the Scenario drop-down as shown in the following screenshot:

2. Select the ImageResizer-CSharp template as shown in the preceding screenshot.
3. Once you have selected the template, the portal prompts you to choose the following parameters:
   • Name your Function: Provide a meaningful name. For this example, I have provided CropProfilePictures.
   • Azure Blob Storage trigger (image):
     • Path: Provide the path of the container (in our case userprofileimagecontainer) which contains all the blobs that are created by the Queue trigger. CreateProfilePictures in the previous recipe
     • Storage account connection: Select the connection string of the storage account where the container and Blobs are stored

3. • Azure Blob Storage output (imageMedium):
     • Path: Please provide the name of the container where the resized images of size medium 200*200 are to be stored. In this case, userprofileimagecontainer-md.
     • Storage account connection: Select the connection string of the storage account where the Blobs are stored.
   • Azure Blob Storage output (imageSmall):
     • Path: Please provide the name of the container where the resized images of size small 100*100 are to be stored. In this case, userprofileimagecontainer-sm.
     • Storage account connection: Select the connection string of the storage account where the Blobs are stored.

4. Review all the details and click on Create as shown in the following screenshot:

5. Fortunately, the ImageResizer Azure Function template provides most of the necessary code for our requirement of resizing the image. I just made a few minor tweaks. Replace the default code with the following code and the code should be self-explanatory:

        using ImageResizer;
 
        public static void Run(
          Stream image, Stream imageSmall, Stream imageMedium) 
          {
            var imageBuilder = ImageResizer.ImageBuilder.Current;
            var size = imageDimensionsTable[ImageSize.Small];
            imageBuilder.Build(image, imageSmall, new ResizeSettings
             (size.Item1, size.Item2, FitMode.Max, null), false);
            image.Position = 0;
            size = imageDimensionsTable[ImageSize.Medium];
            imageBuilder.Build(image, imageMedium, new ResizeSettings
             (size.Item1, size.Item2, FitMode.Max, null), false);
          }

        public enum ImageSize
        {
            Small, Medium
        }

        private static Dictionary<ImageSize, Tuple<int, int>> 
         imageDimensionsTable = new Dictionary<ImageSize, Tuple<int,         
         int>>()
        {
            { ImageSize.Small, Tuple.Create(100, 100) },
            { ImageSize.Medium, Tuple.Create(200, 200) }
        };

6. Let's run a test on the RegisterUser function by submitting a sample request with firstname, lastname, and a ProfilePicUrl. I have used the same inputs that we have used in our previous recipes.

7. In the Azure Storage Explorer, I can see two new Blob containers userprofileimagecontainer-md and userprofileimagecontainer-sm as shown in the following screenshot:

8. I can even view the corresponding cropped versions in each of those containers. Following are the three versions of the image that we have used as input:

We have created a new function using one of the samples named ImageResizer that the Azure Function template provides.

The ImageResizer template takes input from userprofileimagecontainer Blob container where the original Blobs reside. Whenever a new Blob is created in the userprofileimagecontainer Blob, the function will create two resized versions in each of the userprofileimagecontainer-md and userprofileimagecontainer-sm containers automatically.

Following is a simple diagram that shows how the execution of the functions is triggered like a chain:

• • The Building a backend Web API using HTTP triggers recipe
  • The Persisting employee details using Azure Storage table output bindings recipe
  • The Saving profile picture path to Azure Storage Queues using Queue output bindings recipe
  • The Storing the image in Azure Blob storage recipe