Microsoft Azure Functions Fundamentals

1. Introduction: Hello and welcome to my course, Microsoft Azure Functions developing serverless solutions. I'm your instructor for war Williams and I've been a software engineer, lecture and code practitioner for the past decade. In this course, we're going to be focusing on Azure functions and how they work. We looked at how to create them, test them, deploy them. We looked at the different function trigger types and whole weekend MY them interact with external services, also provisioned by Azure and otherwise, going through all of these concepts, we will also interact with different development tools, all geared towards helping us to be as productive as possible. Tools like Visual Studio, visual Studio Code as 0 Functions, core tools. We looked at the storage emulator and the cosmos DB emulator. Know you're probably a little fuzzy as to what Azure functions are. Their phone shuns is one of the quickest and easiest ways to get your code running in Azure while maintaining a cost effective and serverless model. In other words, you can write an application without having to worry about the infrastructure or any servers or anything. And he can deploy it to the Cloud and have something up and running within a matter of minutes. It's also supports many languages. So even though we will be focusing on dotnet and C-Sharp, you can also use JavaScript, TypeScript, Python, Java, and other languages right out of the box. Now throughout the course, we will be focusing on as your functions, but we will interact with other Azure technologies, such as Azure Storage, azure AD, azure Cosmos DB, Application Insights and app services. Let me say again how excited I am to have you in this course and I can't wait to get started. See you soon. 2. Introduction to Azure Functions: In this section, we're going to be demystifying some of the mysteries that surround the concepts of Azure Function. One, I'm assuming that you already have an active as era cones and that you have enough credits to get through the activities related to this course. Some of the things will cost money, but we're not at the money costing part just yet. I just want us to get comfortable with Azure portal works how we get to interact with functions and some of the concepts and keywords surrounding their operations. So stick around. We're going to get to understand exactly what a 0 functions are and look at the various ways that they can be used. 3. Create a Function App: Alright, so let's get right into understanding a 0 functions. Now at this point I'm making a few assumptions. One, that you already have an active Microsoft Azure cones and you do have your credit card setup or you have the free credits available as some of the activities for the duration of this course will cost some money. But we're not really at the money-making part yet or the money spending part. It we're just looking at the fundamental concepts. I'm already logged into my Azure account. You can get to the Azure portal by going to portal.azure.com and login with that account that you had used to create your your account. And then you'll see a dashboard. It may or may not look like mine actually put some time and effort into customizing this one. You see here I have some resources that are deployed from another course. And you will see, you will have the ability to customize this. You can always go to New Dashboard, blank dashboard, or you can just edit this one and you can shift taus all our own and customize them as you wish. No, we are going to jump over to the functions up or at least the service listing and create a function of a function up Israeli an App Service. So you would have seen the keyword serverless being advertised when you see a 0 functions. It is serverless because there's no physical server involved. It's a service that is being offered that allows us to host functions on. So I can actually just go to All Services here. Then I can filter and say function. And then from here I will see that I can create a function up. When I do that, I can click Create, Create Function up. To proceed, I'm going to go through a wizard. So firstly, what is your subscripts? So when you have different subscriptions, you're probably under trial. You're probably on pay-as-you-go dev slash tests like I am. You are probably on a special one for your company or for your organization, whatever it is you choose the one that you're going to be using. I only have one. You would want to create a new resource group. So you can go ahead and hit Create New. And I'll see function, RG, function dash, dash RG. Click. Okay. What do we want this function up named to be? I'm going to name itself with the course which is Azure, Functions, dash fundamentals, know when you're doing that, as you can see, it's actually going to be a public facing address. So what you can do to make it more customized in case that's not available. You can also put dash and your initials or your name or a part of your name just to make it a bit more unique. That is going to be on public, publicly accessible URL that would allow us to browse to the functions that will be deployed our access the functions that are deployed in the central space from anywhere in the world. I'm sure we all know how URLs work written. Alright, so next we want to say published, do we want code or a Docker container for now, I'm going to go with code. What runtime stack do we want? A 0 function support many different languages. We see it has support for the dotnet framework or platform. We can use NodeJS, Python, Java, polar, Show, custom, and there are other languages being added. Very irregularity for this course over we're going to be focusing on dotnet. So we'll choose that then we choose the version. Dotnet 60 is the latest and the greatest. So we will proceed with that. At the time we were doing this course, they may have a later version. You can feel free to use that also. If you're not up to dotnet six, you can also choose dotnet Core 3.1 might be necessary if you don't have Visual Studio 2022 or higher at the time you are doing this course, that's fine, but I will be showing with done at six. And then your reason would be the datacenter closest to your geographical location. If you're in Australia, you want to choose the one in Australia that's closest to you. I'm in the Caribbean. The best one for me is usually East US. And I'll choose a stool that is newer datacenter to East US. After doing all of that, we can just click through and see the other potential sitting so we can set up a storage account because we have to create a link to a storage account for blobs and queues and tables storage, any form of interruption with those. We can choose an operating system. I'll leave this as Windows. We can also choose a plan type, not the plan type is going to directly affect the costing and hold the resources are spooled to support your functions up. So by clicking, learn more, you can grow so the documentation, so the consumption plan will scale automatically and you only pay for compute resources when your functions are running. The downside to this is that they will actually turn off or D allocate the resources after a period of inactivity. And then when you need them again, there was pulled them up, but this may take time, so it's like it's not really always onboard then it will kind of torn down so that it saves you money in the long run, however, if it's something that you always need, then this might not necessarily be the best plan for you right? Now. The premium plan automatically skills based on demand and there is no idle periods, so it's always scaling and it's always on when you need it, it is there at all times. Another thing to note also is the execution time. So the consumption plan basically has a limit of, if I'm not mistaken, 20 minutes. So that means our function will stop running off the 20 minutes. Whereas the premium plan, it will not have that time limit. So it will run longer than the maximum execution time given to us by the consumption plan. Like I said, you have to know what is your contexts and what you need from it. The dedicated plan will actually give you that App Service plan kind of runtime, meaning it will have dedicated resources for your AP function, your function. And you won't have to guess and spell to say, Okay, I'm overspending on the spending. It will scale automatically and it's far more predictive if you know that you're going to need functions and you're going to be doing this long-term. You probably want to use the dedicated plant because then you can actually properly estimate what your costs will be at a certain kind of load. You can choose the one that you need according to your context for this course over, we're going to go ahead with the salt consumption plan, which is told it as the service plan. Next we go over to the networking side. There isn't much or anything to configure here because pretty much it is already configured to be public facing to the Internet. So we don't really need to do any virtual network adjustments here, we can go over to monitoring where we can turn on Application Insights. And that allows us to shoot logs and monitor the functions while they're running the runtime and the errors that they're getting all of those. So I would recommend that you enable that. It does cost a little bit of money, but I think that it's worth it in the long run. And then tags, those are good for creating a naming resources, especially in a setting where you may have many resources are fairly similar nature's serving different purposes. So tags can help to isolate them and help you with building and resource allocation in the long run. But we can skip that for null. And then we go to Review and Create. Once we verify that everything is what we want, we can just hit Create. When that process is completed, you can always just click Go to resource. Once on, Go to resource. If you've ever deployed an App Service, it's going to have a very similar dashboard verse, similar kind of feel. If not, then that's fine. This is the dashboard that you're getting. So you'll see here that he gets some charts, some metrics on runtime notifications. You can do access control to see which user in your organization or in your Azure AD can do what in this particular resource, we can configure security alerts for any intrusion or potential intrusions. We can set up events. And then of course, the reason we're here, we can start creating functions. Someone to leave it there for. No, we just looked at how you create a function up and the different types of hosting models, the different pricing models that are available to you. And of course, context determines the decisions that you make. We're going to go through some concepts in the upcoming lessons. 4. Triggers and Bindings: All right guys, In this lesson we're going to be looking at triggers and bindings, and we're not doing any coding just yet. We're still just exploring in the dashboard to see what swats, what the lingo means and how we can connect the dots between what our needs are and watch the Azure Functions will provide it to us. We already created our function up and once again, that's not going to UPS service that has been to host all of the functions that need to be executed. So we can get there by going to this URL. And if I click it, you'll see that it says, my functions up is up and running. Of course there are no phoneticians for me to browse through, so that's fine. If I go down to that section that says functions and click on functions, then it allows me to see all the functions that I would have, the trigger and the status. So if I click Create, it brings up this bleed that then allows me to define the development environment. You see here that we have a few options which we will be looking at later on. And then they ask us to select a template. This is pretty good and you can see that you have quite a few options in that template section. The trigger represents the action that makes the function execute, because a function is just going to sit there. We all, at this point, I'm assuming our programmers, so we all know that functions need to be in a program and they need to be called in order to be executed. No, The thing with Azure functions is that they are actually more like standalone functions, but they have triggers and a trigger is what will prompt for its execution. So if we look on the needs that templates column, you'll see the different triggered types that are available under these templates. So you have the HTTP trigger, meaning a function that will be run whenever it receives an HTTP request and it will respond. So if you know about APIs, that's all an API is. You make an HTTP called maybe include data, maybe you don't. But based on that call, the API will respond according to how it understood the call relative to the data it may or may not have been given. So that's an HTTP triggered function. As your function, you have timer triggers, meaning you can put it on a timer, execute every five minutes, execute once a day, etc. So, you know, you have that Windows Scheduler or you have those Cron jobs in Linux, you can use timer triggers in the Cloud space to do those kinds of executions in the background. I'm not going to go through and explain what every single type is, but you can always just go through and read the descriptions, right? So while a Zuora lawsuits create stuff, it's actually educating you so you can make the court decision you have as your queue stores triggers, you have Service Bus triggers, topic or cue triggers based on which one you're using for the Service Bus, you have Cosmos DB triggers, right? And then in the same breath you have what you call bindings. Binding is basically you're subscribing to a non-function related feature or features sit and reading or writing data. So you have input and output bindings. So when we say that, I'm just giving you this scenario so you can appreciate what I'm seeing. If you have a timer trigger that is supposed to take an image alt off the Blob Storage or take a message off the queue and put something in Blob Storage. So let's say you had an easier form shown that was service most triggered. Know this phone was shown is going to have a binding to the queue and it's going to have a binding to the blob. This is going to allow you to be able to subscribe to these two a third-party services I want to set third-party I mean, they're not exactly related to the function. They are stand-alone Azure services. So the input binding would be, I'm reading from the queue. The output binding would be, I'm writing to the Blob Storage. I don't want to have to write if it took me 50 lines of code in irregular program to consumer blob and connect to it and know exactly where I'm going. It will probably take three to five lines because of that bindings feature. That's a very powerful feature, allows you to cut out a lot of the boilerplate code and just get straight to the point and implement the business logic. So it's very useful implementing functions or functionality rather that you need like snap judgments, you know exactly what you want. You don't have time to write too much code. That's what these trigger templates bring to the table. And the ability to just bind the function to an external resource makes it very easy to deploy an app that touches many different parts of a 0. That's really all I wanted to discuss here. We're still not creating a function just yet. We're still warming up to the dashboard and lingo. When we come back, we'll be talking about the development tools that are available to us even though we won't go too much in depth. What we've just discussed, what the possibilities are. 5. Development Tools: All right, so the last time we're here, we're looking at the different types of triggers that are Azure Functions might have. No, let us take the step further and actually look at the tooling that would surround us, actually creating an interacting. We'd said functions, no upsilon one, it would be the most obvious option which is to use the portal because the ports, that is a very, very powerful tool. It sits in the browser. It takes a very minimal resources on your actual machine. So this would probably be the feature or the option rather, not a lot of persons might end up choosing when you go to Create Function, the first one there is developing portal. You can choose your template and then hit Create after you give it a name, let us say test the portal. Http function does something so we know exactly what it is authorization level that has to do with the levels of authorization that we're going to be giving persons in order to interact with this function. So I'm just gonna say a function, you just interact as a functional level versus anonymous versus admin. So what happens is that when you use function, you need a full notion key. When you use admin, you need a master key and you can actually manage those keys underneath the key management section here. So if I just say anonymous, then no keys needed. So I'm just going to do this for this, we just hit Create and that gives us smaller dashboard. We can go and look at the code and execute and tests right here in the portal. And like I said, I'm not getting too much into the code, right? No, I'm not going to explain what this is. I'm just showing you that this is a text editor right here in the browser so you can save it, you can test and run it right here. You can do everything you need to do. We can get the function URL. All of that can happen right here in the portal. So that's option a for your tool to write a function. No option B would be Visual Studio and Visual Studio here. The one I'm using is Visual Studio 2022 Community Edition installer. You'd want to make sure that you have that Azure development workload selected. So amongst all the other workloads that you may need, make sure you have included the Azure development workflow. And then once you have that workload, when you are creating our projects, you will be a load or you will be given the option to create an a 0 functions projects. This would be, this would come in handy if you have a whole development project, you're working on a team. You want source control, management and all these things. And you have that solution with a number of functions that you could deploy. Tool the function up. And if you use Visual Studio irregularity, then I don't need to marketed to you as being a very good development tool. Another option you would have would be Visual Studio Code. And you can always go to the extensions and installed as your tools. So as 0 tools will come with a bunch of extensions for all sorts of Azure development, including a 0 functions. If you don't want the entire toolset, of course, you can look specifically for the Azure Functions, but why not just install all the tools? Now this option is perfect for you if you are using Linux or Mac, because Visual Studio code is open source and cross-platform. So it's not limited to the Windows PC like Visual Studio might be. It would be a good alternative once again to have the code on your machine and source control and so on, outside of it being in the portal. Now the next option, which admittedly, I have never attempted to use, but it is there. So we will, we will discuss it would be the any editor plus core tools, right? So we can go ahead and use Node.JS and install of these libraries. And then they tell us how we can start to tell we can test it or we can create a project up to how you can publish it. So if no GS is your preference, then you can go ahead and follow these instructions and go through that. But like I said, we're not doing any coding in this particular lesson, I'm actually going to delete this function because I don't need it. Beyond know. When we come back, we will continue discussing some more concepts and then we will close out this section and get right into some of the fun that we came here for. 6. Section Review - Azure Function Introduction: All right, so let's look at those as sexual notes. Let us review what we accomplished. One, we looked at what it takes to create an a 0 function up. And we discussed the fact that the function is a single unit of deployment. It's an App Service that is going to be hosting a number of functions. So a function can be created and it can have different templates, are different triggers. So the template is just the predefined cold, but you can always deploy any kind of template and read the code accordingly. But we have different triggers that will set off the functions. We looked at an HTTP triggered function already. We have timer will have q, we have Service Bus triggered. We have a number of them that we can use, durable and MQ. Later on we will look at what the Durable Functions really mean. But the point is that we have all of these options available to us. We also looked at some of the development tools that we have available to us. We will be using Visual Studio or I will be using Visual Studio. But mostly, if not, everything that I do can be accomplished in the portal as well as using the other tools if you need to use them. When we come back, we're going to delve into interacting with functions using the portal a bit more and we're going to get our hands dirty with some code, hold to a test or functional and sold to look at the execution and actually see what it takes to create and deploy some Azure Functions. 7. Azure Functions in Azure Web Portal: All right guys, welcome back. In this lesson, we're going to get right into the fund and we'll be creating a few Azure functions right here in the web portal. So we're going to be looking at examples of HTTP triggered film showings. We'd be looking at timer triggered functions. And we're going to look at one with Blob Storage, notwithstanding that you can use those same steps and deploy any kind of function that you need according to watch you need as the trigger will be focusing on those three examples. And we'll also be looking at how you can test. Well, we can look at the logs that the things that you can do with the functions just to make sure that you are doing it properly and covering all the bases. So stick around. We have a lot of fun overhead. 8. Create an HTTP Request Triggered Function: Alright, so let's start off by creating an HTTP requests triggered function. And we did that already. So we're just going to replay our steps. We're in the function up and then we go down two functions, then we click Create up top. And then that gives us the list. And from that list we want the HTTP trigger. You click that, you give it a function name. I'm just going to name it HTTP, request hyphen funk, or a function. Let me just write that out. And then here for the authorization of Endo skimmed over last time, but I'm going to go much more slowly this time and explain in more detail when we, when we choose the authorization level for function, what's going to happen is that any, any device, any color to this function needs to provide a key that is specific to this function. So by creating a function and we get a key, and this key must be present in every requests to this function in order for it to actually be triggered. When we say anonymous, it's open, anybody can access it. There is no need for our key or any additional security constraints. Then when we use admin, it means that the admin who was a master key that would allow the admin to execute any function whatsoever, any function at all, but only at the admin level cannot be executed. I'm going to leave this at the function level for null and I'm going to hit Create. Now that our function has been created, we can go down to the code and test bleed. Once we're in this blade is going to show us some boilerplate function code. A lot of things that are defined for us. You'll notice that the file extension here is dot CSS. So based on the template that you're using, meaning are using JavaScript or using C-sharp, etc. Then that file extension will of course be different. And if you drop that down, you'll see that you also have a read me file dot MD or for Mark don't similar to what you would have on GitHub or a 0 DevOps for documentation purposes. And you also have function dot JSON. And this function dot JSON file basically just has configurations for different bindings and different metadata about the function. So here you can see it's specific that we're using the HTTP trigger and it's accepting the methods of GET and post. So if we wanted to do put and maybe patch and delete, etc, you just extend that whole section. There are four methods. There are times when you may need to modify this file, but for now we don't necessarily need to modify this. So we'll just leave that alone and let us go back to our Randolph CS6. And what we're going to do here is look through the code. So it's really simple code where using a few libraries. And here we're including third-party library that you would have probably out to get with NuGet if you're doing irregular dotnet application. But we're using some libraries here. And then we have a method that returns an eye action result. So if you've ever worked with APIs, you'd be familiar are even our regular MVC up. You'd be familiar with the eye action result. Note that HTTP triggered function is usually very useful for pretending or not pretending, but for acting like an API. And it's also very useful for responding to webhooks. So anything that is HTTP traffic in HTTP triggered function, generally speaking, is equipped to handle that kind of operation. And so you could actually build out a full API or full application logic using a suite of HTTP requests triggered functions, each one responding to the specific request accordingly. Here, inside of our method, we see that we get to parameter is GTP request. I will call me at Rick, and we're also putting in an instance of our logger. So we're logging information first-line where they're seeing this was a process. It was called we are processing our request. Then we're getting from the request a query string with the name name. So we're getting the value from the query string with the name name. And we're storing it inside of this variable called name. And then we're seeing the request body process that request body, please. I'm looking that request body if it's JSON or stream it to JSON, look inside of it and see if you can find name. Here it is saying that either you're going to find the query string with the name name in the query string in the URL itself. Or you can look in the request body in case it may have been a post request and look for some data that has the name name, please give me that body. That's all that that's just doing. It could've been on either are because usually you send either upwards or I get what's here. The code is just showing you that this is where you get a query string value and this is where it gets something from a pollster requests. That's all that's really happening. Then it's going to formulate a response message. And it's good to say if this name variable is not null or empty, or sorry, if it is smaller, empty, then give them an error C and that you need to pause it either in the core string or in the request body. And if it is not dollar empty, then we're going to print Hello name. This was executed successfully. Then we return, okay, which is a 200 results if you're familiar with HTTP response codes, okay? The OK function gives back a 200 response code as well as that message. All right, that was a tour of the code. Not going to modify it. This is going to test it. So let's test and run. We see here firstly is going to say OK and testing, are you doing a post or a get method call? Secondly, what kind of key, right, so I can use the master key monster because in the portal we are the admins and a master key can do everything anyway. Well, you could just test it with a function key, since that is what we deployed it to use. You can add a parameter here or you can change the body here. So I'm going to my name inside of the body. That is a post request. So when I click Run, then I'm going to get the ALU output alpha 200 response, hello, the name that I entered. This was triggered successfully. So that's a good test. Now let's jump back over to input and change the parameters. So I'm doing a GET request this time and I'm putting that query and I'm calling the variable name. So the parameters in the name and the value is still my name. And I'll just say dash get. And I don't know if you're experiencing this, but while using the portal, I have to press a button and then wait for the character to appear. So that's one of the reasons that don't really like working in the portal directly like this, at least for this kind of operation. It's a little slow for me, but you can let me know if you're experiencing that also. But when I click Run, then we get the same results. This time it was getting it from the query string as opposed to the request body like it did for the post. Now if I miss name this query parameter and then run, then we get to our error. This HTTP triggered foam Sean executed successfully. But please pass in data as it is expected. That is pretty much hold. The HTTP triggered function works. I request comes in with data. We write code to parse through the data, manipulate it to whatever we need to do with it. And then maybe we send back, we write a log so that we can actually truck the execution path. But we also just let them know the kind of response. So I could actually respond NC, if the query parameter was not null, then return a new maybe bad request object results, right? Something like that. That is pretty much holdout works. No, I can actually get the function URL by clicking this button up top. Then it will say, give me the default. Well, which link do you want? So it's actually going to include the key on that URL. Do I want the master key or do I want a function key? So I went to leave it on default. And if we take a look at that URL, and I'm just going to use Notepad Plus Plus. Here we see that we have the URL to the function up. And then it says slash API, and then the name of the function. And then code is equal to, and then this hushed string, which is that key, That's the function key. So if I needed to execute this from any kind of application that I am writing, then I have to execute it again. So you are looking at minimum like this. Without that code, it would never reach the destination. He would just reject the request altogether. If we go through the other pages that are available to us from this screen, we can go to integration which allows us to edit certain things. So we candidate trigger, we can edit the function, we can modify the inputs and the outputs are some mini flowchart, just to say that this is where it's starting, this is the process and this is what is expected. We can go down to the monitor, which basically shows us all of the executions and homeowner for them were successful, how many of them failed? Then we can go to the function keys, which will show us our keys. We can always click to view the value because we're using the function key, that is our function key if you ever need to retrieve it, if you need to renew it, reset it, etc, whatever you need to do, that is pretty much it for setting up an HTTP triggered function. 9. Create a Timer Triggered Function: The next type of function that we're going to be looking at is a timer triggered function. So this dumped onto functions. Go ahead and click Create, and this time we want timer triggered. So we clicked Timer Trigger and we can give it a name, someone to sit timer triggered function. Then we can set up the schedule. The schedule is based on a chronic expression. So if you're familiar with Linux, um, then chronic expressions are just way you represent the interval that you want something to be repeated on. So you can see here that it's in the form of a second, then minute, then all we're so each asterisk basically represents that that unit of time. Pretty much here. It's seeing 0 seconds, but five-minutes. So if I change this to one, then that would be like ever minutes as opposed to over five minutes. And then that asterisk would represent the lower than the d and then the day of week or the month and day of week. So that's this all you would really schedule that it may take a bit of getting used to if you're not familiar with how it is, but it's a very powerful constructs to understand and appreciate. So let's go ahead and create this timer triggered function. Once that's created, we can jump down to code and test. And you see here that it is a relatively simple bit of code. We have a function called Ron, and it has a parameter called my timer, which has all of the things you may need surrounding an object that allows it to glean. I'm details about the timer being used and it has the logging library also. It will be logging to the information or making an information logger other each time it is executed showing the timestamp. If I look in the logs every minute, you are going to be seeing this populate, so I'm going to leave it so minutes and then show you the aftermath. Alright, So I waited for two minutes. And there you can see the execution for minute one and the execution for the second time. All right? That's pretty much what the timer service brings to you. I'm sure you're thinking about different applications that this might've, it could be in an application where you need to do a clean up every morning or every night, or you need batch processing to hop in everyday at a certain time. You have a number of applications for timer, timer triggered functions. In this situation do I don't need it running in the background. If you don't want it running in the background, you can always just disabled. And we already looked at how we can delete a function. So I just disabled this one onto maybe in the future when we use it again, but that's really it for a whole. You create timer triggered functions and how they work. 10. Create a Blob Triggered Function: All right, next up, let's look at what it takes to create a blob triggered function. So let's create in the blade, we're going to look for blob. As soon as it loads up. There we go. So as your Blob Storage trigger, right. So you see there are a number of them and I encourage you to actually try them out. You might have useful them outside of these demonstrations that I'm making. So I encourage you to go ahead and try see what the quarter looks like, see what the binding code looks like. But here when we do the Blob Storage trigger, I can change the name so that's blob triggered function. And then our path is going to be, basically the path is going to be where should I look? Slash, end up binding expression, what should I look for? So here I'm saying I'm looking for the name of whatever is in the blob. That's a binding expression. Give me the name so it allows me to access that thread. And then the storage account connection, it's pretty much going to ask which storage icon should I use? So it's going to be defaulted to one already associated with the AP that the function is being hosted in. So I'll go ahead and hit Create. And for this one, let's go to code and test, where we will see that we have information coming in from one, the stream which is being called my blob. And then the name which is that binding expression that we set up. And then we have the logger. Of course, this is sitting down and watching that blob. Now let's switch over to the function dot JSON and look again. So we have the name of the type Blob Trigger, we have the path, samples, dash work items, slash the name. So what we need to do is have a blob that exists with that name. If I look in my storage accounts and to get to the storage, I'll go and just in case you're not familiar, you can go to the dashboard, you may see it under all resources. If not, then you can just go to storage accounts. And then you can click storage accounts. This no-shows all of your containers. So these containers are really the species that the files will be stored in. I don't want to have a container by that name. Whatever name you I'd sit there some dashed work items was the default. We didn't change that. What CPI change it, whatever name you put there, you have to make sure that you have a blob that exists with that name. So I'm going to create that container and just go ahead and do that and know that it exists. I can start adding files. So I'm just going to upload a test file that I have on my computer. So I'll just click Upload, select the file. And sometimes this might be faster if you have the Storage Explorer. So that's actually a third-party tool that if you wish, you could install Microsoft Azure Storage Explorer that allows you to handle all of his storage accounts and everything from your computer, but that's a conversation for another time. Split the file and then choose my test file and upload it. And you see that I uploaded it in tab. If I go back over to the function and look at the log you see here it is logging that it sees a new file and that is the path. You see the name there, blob dash test.txt. You see the creative time insertion time on all these things. That's just how you can sit down and watch the storage ear to see. Maybe you wanted to compress images as soon as they're uploaded. Maybe you want to move them around or renamed them. There are a number of things you can do with something like this. Once again, your contexts will always determine your application. 11. Section Review - Different Azure Functions: So where are the end of this section? And we're just going to review what we've learned. One in case you missed it the first time we did it again, and we did it three more times. We know know how to create a new function. So inside of our function up, we would create and then we choose from the list, the one, the templates that best suits what we need. It's better to just create it from the ten feet as opposed to create one and then customize it for another. But let us just work with what it gives us as templates. And we looked at the HTTP trigger, we looked at the Timer Trigger and we looked at the Blob Storage trigger. However, based on the different resources and services that you may end up interacting with, you may just start off with the appropriate template from the get-go. In looking at our HTTP requests to focus on, we saw that we get a mini dashboard that allows us to see sort of the executions which ones failed, how many times we've got successful cause, maybe 400 calls, 500 clause, etc. All of these, um, are readily available to us. We can get the function URL, which we established has the code or the key attached to it, which allows us to plug it into any application. We also discussed that when creating the function, at least the HTTP function, we can create it as anonymous or for anonymous access for a function level security or mosque deliver security or admin level security. We looked at all of that. This one is restricted out of function level as it stands. The code plus test. We can look at the code that we need to modify. Of course, what this one does is it looks in both the query string and the request body based on the type of requests that we're doing, we wouldn't use one and not the other maybe or both. It depends on your context. We can always test on her own from this interface and we can view the logs right here to see what is happening based on our log information that we write in our code. So this is the console era gives us all of the data. Accordingly. Know, when we go to integration, it gives us a nice little workflow kind of screen that allows us to modify some of the bindings and we will be looking at bindings later on. So we have input bindings and output bindings. We have the trigger that we can modify. We can jump back up to the code. If we look in monitor, we see here that we get access to the application insights. Because remember we enabled that when we created the function up. So we can see all the invocations. Each time it was executed, what the result was, how long it took, was a successful execution or not. We can also look in the logs for more detailed logs of all that happened or is happening with the app. At that time. We can also go down to function keys and we can create a new function key if we want, we can revoke one, we can renew. If we need to have complete control over how this function works. And all of that is just the HTTP request function. But we also saw that while similar, each function type has its own nuances based on what it is there for. The timer triggered one, it will do something every so often based on the Cron expression that we specified. The Blob one was going on watch us space for our file to be dropped there and then it can process it. So all of these things are possible with all of that done, we've completed this section. When we come back, we're going to be looking at using Visual Studio to create a function up. And we will see what the differences are between doing it in the portal and using Visual Studio on our machine. 12. Azure Functions in Visual Studio: Alright, so in this section we're going to be creating some functions are a function using Visual Studio. What I've done here is I've jumped over into Porto, creates, went to functions and then I selected the Visual Studio from the drop-down list. And they outline everything that you need in order to get that up and running. They're still mentioned in Visual Studio 2019. So if that's what you're working with, then no problem you can follow along. But I will be using Visual Studio 2022 for this set of activities. You can just follow along these instructions. And really and truly that's pretty much what I went to be going through with you. If you want to just read that and skip the rest of these lessons, then that's up to you. But for now, I will continue and we'll just in the next video, jump over to Visual Studio and get started. 13. Create a Function App: All right, so at this point I'm assuming that you've already installed Visual Studio and you've already set up your Azure workload, like we discussed at the beginning of this course. To get started, we're going to jump over and create a new project. And then you can filter. You can always just search up period of templates for as your functions. Or if you already did at once. At some point in time, it would be to the side. Here. I'm just going to see a 0 functions Fundamentals. That's the name of the project and the solution. And then I just hit Create. At that point they're going to ask me a few questions. One, what kind of template would I like? So similar to the portal where we could choose a template, they're asking me, Do you want a service both trigger, etc, etc. You'll see in the templates. This time I'm going to go with empty those so that we can do a little exploration. So I'm going to say empty and then I can use a storage emulator. So if you are doing this without maybe having direct access to the Azure subscription, or you haven't gotten that far with Azure as yet, you could actually just use a storage emulator where we'll pretend that you have a storage of cones bought off course with some limitations. So I'm just going to proceed with all of that and go ahead and hit Create. So once that project is created, we can look at some of the files that are available to us. So I wanted to start off with the project files. So in dotnet, things since.net Core three-point one O boards, when you click on the CS approach file, you can actually see the XML file behind it. You can always right-click and go to Edit Project file and you will get this view also. This file That's a little bigger. This file really just outlines the font that we're targeting, the dotnet seeks framework and we're using Azure Functions version for at the time of this recording, this is the latest and greatest. And when we are going to be working, we have to reference the package Microsoft.net.SDK dot functions. And here it's talking about two files that are given to us host dot JSON and local dot settings.js SON. And they're letting us know that they are going to copy them to the output directory for the local settings.js on each never go to a published directory. And that's really because it's a local file, but we will see why in a few. So if we jump to the host dot JSON file, it's just a configuration file. It's saying OK version 2, we will see what that stands for in a few, but it's also has some configurations are on logging where it's seeing it should use the Application Insights, which is what we know we enabled in our function up. That's really all that is going on there. When we look at the local lot settings, JSON, it's letting us know that it's not encrypted and it has some values. So the a 0 Web Jobs storage, we know that that is actually supposed to be the storage account that's associated with a function up. But because we're using the development storage that is set to true, so that's the value it's seeing. Just pretend that it exists pretty much. And then it is giving us the oceans work or runtime which is set to dotnet, which could be something as based on the type of language that's being used. But because we're using C-sharp, we need a dotnet run time. So that is by default set to dotnet. So that's a basic tour of this functions up. And the two files that we got. When we come back, we'll actually start writing some code to write a function. 14. Test Function in Visual Studio: All right, so now that we have our project up and running, What we don't have is an actual function. Alright, so I'm going to right-click on the project, go to Add, and then I went to see new Azure function. When I do that, I will choose the name of this function and I'll just leave the defaults function one-point CS don't want this is not dot CSS this time, but I'll leave the default name. Go ahead and hit Add. I know this dialog window asks us which template we would like. So we've got that first when you are creating a project, we went with an empty one, but its buck. For each function that we're going to create, we're seeing the same templates that would have seen in the portal. And each of them would be asking us the same configuration questions that we would have seen if we were creating it in the portal with a timer was still get to set up the schedule with the blob, we still get to set up the connection. Well, this one is a bit different than what we saw before, where we sit up to set up the path and a number of things. Well, I'm going to go with a path of least resistance. Let's go with the HTTP trigger and I'm going to choose an anonymous one this time. And let's go ahead and add. Once we do that, we get this function file. And I'll notice all we got was that CSV file dots, the dots CS6 five. We also don't have a function that JSON file. So what happens is that when we use Visual Studio to create our functions, It's actually going to compile it down into a DLL. And if you've been using dotnet long enough, you know what a DLL is. And then that DLL, we'll have all the bits of information that the portal needs to use once it's deployed to ascertain any configurations and anything else I wrote it. So it's going to be precompiled by Visual Studio and just deployed to Azure as opposed to meet in a 0 and then compiled in a 0 and then used in Azure, like what we're doing in the portal. So you're probably more familiar space when you're writing your code and you have better access to intelligence and resources here in Visual Studio, we can also handle debugging much better because we could actually just run it in debug mode and set our breakpoints just like we would with any other kind of application that we will be building inside of Visual Studio. Here we have a static class and then we have the function name. And then we have static method called run, which knows it's an HTTP trigger. It knows it's authorization level is anonymous, which I can always change here to be function both of course, you'd have to have some more configuration around that with the key, etc. But that's anonymous. We have the different types of requests that are accepted. So GET and post. And then we can define a root and a number of things that are available to us right here. Ultimately though, the code or the code in the body of the function itself is pretty much like what we had seen previously. If we were to run this and I'm just going to hit F5 and I'm going to set a breakpoint just to show you that debugging those work. If I hit F5, it's going to Visual Studio. Visual Studio things. If it's your first time, you may see it installing some tooling as it goes along. But here we see it's open running. So it's letting us know that in order to get to this function, we have to send a request through this URL. I'm going to borrow that URL and I'm going to use a tool that we use for EPA testing called Postman. So if you don't have post-money, can always go ahead and install it. It's pretty easy to find POSTs ME n, but it's a good tool to have just descend random HTTP requests when you are testing. So I'm just going to go ahead and put in the URL. I'm going to leave it as I get. And for the parameters I'm going to put in that name parameter. And I went to see test from postman as the value. Then when I send, it's going to connect to our app that's in run time, IACUC taking the breakpoints. So note that it's setting a breakpoint, I can actually interrogate my objects. So let's look at the request itself. The request object has everything about the HTTP requests. It has a myth that was being used. The path, everything that I need to know pretty much if I don't inquiry, I'm going to see that the okayed found one in the list and it has the key name and the value tests adjustment there. I can ascertain that it is supposed to get the name right. So if I just hit F5, I can remove this breakpoint know, and I just hit F5 for it to continue. Then the console window came up when you went to the debug mode. But if I go back to Postman, which is the tool that made the request, then you're going to see it responded with a 200. Okay, hello tests from postman. This HTTP triggered function has executed successfully. That's just hold local testing can happen, no. So we can actually use tools like Postman and any other tool that you might have, maybe Fiddler or even your browser. And try and do these kinds of operations. We can debug know, and we can better interrogate our application or a function, sorry, code as we're building out or business rules for node, that's really it. Now we know how to create a function using Visual Studio Hall to test it and hold to integrate with third-party tools in your browser to trigger HTTP or requests. Know, as we go along, we'll get a bit more complicated. But for now, that's it for how we create a function using Visual Studio. 15. Section Review - Functions in Visual Studio: All right guys, so that's really it for this section, we looked at how we can set up Visual Studio 2022 or 2019 with the Azure related to sets. So we installed as your workload and that allows us to create our functions. And this is really just a function up. So you see here, each function that we would create in this project would be a standalone file that is a function. When we deploy this, which we will be doing later on. This is just getting our feet wet. But when we deploy this entire project, it's really deploying it as a function up with different functions. If we have a number of things you want to accomplish and a function per task, then we have one project with each function for task, one setup configurations governing whole all of these functions interact with each other and probably any other service that they need to and any third-party service at that one deployment is done. When we come back, we'll be looking at how you can use Visual Studio Code to develop your Azure Functions. 16. Using Azure Function Core Tools: All right guys, so let's get right into it for this section where we'll be discussing the Azure functions using Visual Studio Code as well as the core tools. So I'm on a pH here, we're work with a 0 function core tools is the heading so you can get there by, you can easily just Google your foam shedding core tools, and it's the Microsoft documentation Ciceronian, the core tools allow us to develop and test functions locally. So this is a better option than Visual Studio because it is cross-platform. So it creates a runtime on your machine to simulate what the 0 function runtime needs to look like. And then you can develop and test locally, of course, before you publish. Of course, once again, visits to the allows you to do that, but that's not always an option based on your operating system and other limitations that you might have. What we can do with the core tools, if we scroll down, we see the different versions. So right now I'm using version four because I'm using dotnet six bullets, of course, based on whatever stuck you're on or whatever version of dotnet, make sure that you're going to use the correct version of the core tools for installation. We have the Windows, Mac OS, and Linux instructions, right? So based on your operating system, you go ahead and use whichever one you need to. I'm on Windows, I'm going to use the 64-bit one that launches the typical Windows installer that we're used to. And we can just go ahead and hit Next each time and let it install. And while that's happening in the background, I'm just going to point until they can also get to the core tools by GitHub. So you can go to the Azure project and look forward to Azure Functions or Azure accounts rather, Andrew for Azure Function core tools as her Functions core tools. And then that those have installation instructions according to whatever stuck you're on, probably a little bit more information than you would readily see on the Microsoft documentation. So that option also exists. No one says finished, I'm just going to launch my Visual Studio Code. And what I typically do whenever I have a new project to build is I just go ahead and create a new folder. I created and open a new folder somewhere on my machine. And I'm kind of assuming that you're familiar with Visual Studio Code, but if not, what I typically do is just click Open Folder and navigate to where I want its create the folder and then use that folder that I just created. And that's what creates my project. I'm going to do all of my 0 function unrelated operations right here. In this interface. Let me just make this a bit bigger so that we can see it. The next thing that I wanted to do is open a command prompt window. So if you're in Linux, if you're in mock, you might be different from Hawaii, I'm using it, but at the end of the day, we can all agree that command prompt looks very similar across every platform on all the commands that I will be using here will be usable in whichever operating system you're using to verify. Well, firstly, what I did was navigate to the folder where I know my project is. And then I didn't see. And that's where we went to verify that the core tools have been installed successfully. So we get this nice particular untold. And then we have of all the documentation around the different commands that we can run. Here we see unit would create a new function up in the current folder and initialize a Git repo. So let's try that one. So I went through it if UNC space in it. And then it's going to ask me to select the runtime. So of course I wanted to split the dotnet run time since we're using C Sharp and low it, so it's doing its thing. I see here it's populated in the files in the background while it did its thing. No, I did try that same command inside of the terminal that's built into Visual Studio Code. And as you can see, I've got an error. So I was hoping that we could do everything from inside the terminal instead Visual Studio Code. But clearly there are some limitations and that's no problem. So we just have to understand which tool can be used for watts when we need it. Just glance over at the fuzzer were created. We have this thought VS code folder which has an extension file just to make sure that the Azure tools or the 0 function tools are installed, give us the best experience possible in Visual Studio Code. We also have the gitignore. It created that CSV file that we're familiar with and the other two JSON files which were also already familiar with. If I wanted a new function and just, just a note, remember that when we did func init, it's asks us to runtime. So you may not necessarily want a dotnet.net runtime. You may want an order on time or a Python runtime based on your skew towards your preferred programming language and environment. And that's completely fine. The fact is that it supports many different languages. I'm only focusing on dotnet because I have a Windows machine. And if you were using a mock and that's not really an excuse. I'm just saying I'm focusing on dotnet because that's what we're doing for this course. So if I want a new function, I will use that if you didn't see command again and then I would say New. And if you're not sure, you can always go back up and look at the commands that would allow me know to select from the list which new type R, which template I would want for my new function. So I'm going to choose one that we haven't done before, this little Cosmos DB trigger. I can do that. Press Enter, then it's going to say what should the name b. I went to see Cosmos DB triggered function. And then when I press enter, it will then go ahead and do its magic. And in the background you'll then see the appropriate file pop-up in your Visual Studio code. That's sums up whole weekend, use the Azure core tools and the command line to start the process. Now the thing is that you may not be using Visual Studio Code as your editor. You may be using a different editor, but for using the core tools, you can actually do everything right here. And even if you were using Notepad Plus, Plus, you could just go in and edit whatever files are being deposited in the folder once you are running your commands here. So when we come back, we're going to look at managing the project, fusing only Visual Studio Code. 17. Using Visual Studio Code: All right guys, so we're going to be using Visual Studio Code to start up a function and up from scratch. And we'll be looking at the different tools and different nuances that make it different from using the core tools and Visual Studio itself. So that being said, let's get started. The first thing you want to make sure we mentioned this earlier, but I'll mention it again. Make sure that you have the Azure Functions extension installed. Know what I have is the Azure Tools extension with which came with It's an extension pack. So it came with a bunch of Azure related extensions. And I just did that because I do some Azure developments. I prefer to just have them when I'm ready for them then to go hunting at the time. However, if you don't want the whole park, no problem. You can just look for the Azure Functions extension by just searching up here. And that one is just going to give you what you need to make sure that you can do your Azure Function development. Just reading through the documentation is actually quite informative and it does tell you how you can go about installing and setting it up and running and deploying. When the time is right. The emblem in the documentation is slightly different from the emblem that I have on screen, at least at the time of this recording. But that's fine. Guessing that they're going to work very similarly anyway. So I'm just going to click that as your section and then it allows me to view all the functions. I've created a brand new folder, which you can also do if you need to, but let's just work through this. So here I have the option to create a new project, to create a new function when I'm already in our project and to deploy the function up, which would be the whole project. So I'll just start with creative project. Which then asks me, weird, do you want to put this new project so you can go to browse and then creates a new folder wherever it is. I already created a folder called it as your phone, Sean's fundamentals code. It's basically in the same place as the previous one where I was using the core tools will tell us go ahead and select that. And then incidents asked me to select a language. I'm going to mix. It's above it. We've been dealing with C-Sharp all this time, mainly because it was Windows and because I've been using Visual Studio and we'll be using Visual Studio going forward, both in an open source sitting. All of these are options that you can have someone to go with. Typescript does something different, then they want a template for the first function we can skip that will be akin to creating the empty project in Visual Studio and then adding the functions later on. In this situation, I'm just going to start off with an HTTP trigger. So let me do that. I'll leave the default name, press Enter, and I'll make it anonymous. And with all of that donuts creating. And then it finished grids in a project and then it did a bunch of things. And it's initialized a Git repository for me with all of these default files when I go up the project listing, have a VS code folder that has the extensions and initializing files to make sure I have the best experience while using Visual Studio Code, I have a folder for the HTTP trigger which has the function dot js. We already know what this file is far we've seen it before. It also has that index files. So unlike the C-Sharp file, which has a dot CSS file, I chose TypeScript, so it's not TSV file, which is really just an encapsulated JavaScript file. If you do know door, no development or angular and such, then you code TypeScript looks pretty much it's just a JavaScript version of what we already know. It's taking the request as an HTTP request. It does its context, which is the context of the function where logging. And then we go and get the name from either the query string or the body. Then we print out according to if data came back or not. We have some file. We have a function ignore file, I have a gift ignore file, and then we have the host JSON files. So you see here that the runtime is gnawed on the last time when it would be dot in that package file. And we have a TS config file. All right. Those are all the files that we get out of the box. I'm just going to go ahead and run this someone to start debugging. You might see it installing some other stuff on your computer. Of course, having the latest version of Node already installed in this situation is imperative for this to actually run locally because obviously depends on nausea, need to have Node and npm install on your machine. We see here in the console that we have the URL to our function. We don't hold the testis or do we can either use a browser or we could use Postman. So to spoof, request someone to go for a postman, I'm going to put in the new URL, so it's whatever you, or they give you here, slash api slash HTTP trigger one. And then I went to put in the query string for the name. So I went to send that. You see here that the logo and the background went crazy, right? So it got down activity. And of course we got that response as a 200 OK response. That's pretty much it for using Visual Studio code, debugging the formation in Visual Studio Code. And then we can of course deployable. We looked at all of that. We're going to go through the whole development of a whole suite of functions in a few, but we're just taking it step-by-step? No. Just before I move on, you can follow along if you want, but you don't necessarily have to put in the core tools runtime. It is imperative that you know which version of the core tools runtime versus switched version of dotnet you're using. Dotnet six is still a relatively new at the time of this recording. The tooling isn't always going to be laborious and 24 weeks. So here I went through the same steps, once again, creating a new project, but I chose C-sharp this time. And you're seeing here that you can use.net Core three are dotnet F5. I went to choose dotnet F5, and then I went to do an HTTP trigger just the same way. So of course I said all of this open a brand new function by burning folder, sorry. Just repurpose everything here. That is a C-sharp function which looks the same as what we've been seeing. So if I just do a run, I just wanted to ruin and make sure that my runtime is, I'd be quit for building a dotnet project using Visual Studio Code. It was like I said, the tooling might not necessarily be stable. It needs to be at the time, but hopefully by the time you're doing this course, all of those nuances are gone. So yes, it was successful. Here we see the runtime is up and I am successfully debugging my C-sharp beast function. I'm seeing all of that to say that maybe when you spin up with a core tools and you might be in one version and then you try editing it in Visual Studio Code and then try running in debug mode in Visual Studio code, you might not necessarily have the best experience if the tooling and aversions aren't the same, you might end up getting weird errors like missing configurations and such. So I'm just pointing that out in case you do end up like that, It's better to start in one place and do everything in one place. So if you start with Visual Studio code, everything there. Same thing with Visual Studio, etc, etc. With all of that done and said, I'm going to call it quits on this lesson. I hope you learned and field equipped and ready for the project up ahead. 18. Section Review - Azure Functions in Visual Studio Code: All right guys, so forth. This section we looked at alternative tooling to Visual Studio. Visual Studio is a big poll for beast, of course, but it's not always the best or the only option for some persons. We looked at the fact that you can actually get the core tools. You can spin it up from there. You can use a regular text editor and a few commands in the command prompt. And you can at least have a good experience developing your phone. Shown is deploying it and interacting with them. We also looked at Visual Studio Code, which is a very critical sick onto what Visual Studio offers you. And it has excellent tool sets, an intelligence to help you through your development tasks. It allows it to debug just the same we borrowing a few things. It is a very good, like I said, for a second. If you don't have Visual Studio or that's not an option for you. Visual Studio Code is an excellent resource to use when you want to develop your Azure Functions. 19. Azure Function Bindings and Durable (Orchestrator) Functions: Hey guys, welcome back. In this section we're going to be looking at some other features or functions that we probably haven't covered yet. So I went to go a bit more in depth into input and output bindings. We touched on that earlier, but we're going to go through some practical examples. We're also going to look at Durable Functions and art history. And we're going to look at setting up a handlers. So we're going to be using Visual Studio Code for most of these activities. And you will see how to make functions a bit more robust. And the whole weekend let them interact with other services outside of just one trigger or one cause. So stick around. We have quite a few things to go through. 20. Input and Output Bindings: All right guys, so in this lesson we went to be looking at input and output bindings. Input and output bindings for too much referred to literally input and output, a source of input and target for an output. So I already created a brand new project. I went to the Azara tab here. And then I said creates a new project and it shows a node or a TypeScript template for this particular one. So we'll just be using that here. No, I do realize with the tooling that some of the things that we will be going through, you will see different things based on the project template that you're using on the way the tooling works. Remember, a lot of this is brand new. So over the time there might be obeyed same and not to experience what I might point out as a potential gap. However, I am choosing this path of least resistance because we do get a wizard when we want to add an input binding. When we get that wizard and put in the appropriate information on what's going to do is generate a block of configuration code which is global. So no matter which template, even if you don't get the wizard itself, you can still watch if you don't want to follow along our account, follow along based on the template type you've selected. However, when the code is generated and then we review that you can actually stick it into your own In dot JSON file and proceed just a scene. At those points. I will point them all to, you know, another thing that we're going to need for this activity is the Cosmos DB storage emulator. You can get that through an easy Google search, just Google search, a 0 Cosmos DB emulator, go to the Microsoft documentation and you can install that on your Windows PC. Know, if you don't have this option, you can actually go to Azure and provision the Cosmos DB service on your own. And that's pretty simple. The very similar steps to what I will be doing in the emulator. We'll be able to be replicated in the dashboard regardless. So I'll be using the emulator for this activity. So Cosmos DB will serve as our input binding. And then when we're ready for the output part of it, we will provision an Azure table or maybe when we get there we can figure it out based on the activity I'm just playing by ear here. When you have installed the DB emulator and you launch it and started it will come up in the browser looking something like this. And you'd see it tells you it's running, gives you the URI, gives you the primary key, and it gives you the parameter connection string. We're going to need all of that. You can also go to the Explorer. You can create a new database and you can create different containers in there. I'm going to remove this because I wanted to do it from scratch with you guys. So let us start off by getting this connection string and then jump back over to our project. What we need to do is let the project nor about this new connection string jump over to your local dot settings.js JSON file. And here, make sure that the Azure Web Jobs storage is seeing use development storage equals to true. Whatever the form shown in runtime is, we know that what that purpose or the purpose of that is. But let us put in our connection string. So I'm just going to call it Cosmos DB connection. Then I'm going to see colon and then put in the value which is that primary connection string that we just got off the emulator. So let's jump back over to our emulator and create our database. So in the explorer of the new database, and we're going to be, let's say creating a blog. What we'll do is make sure that we have the TIG for provision throughput. We leave it on auto scale. And the cool thing about the emulators that whatever you setup as the settings here, when you are doing the MLA version, sometimes you'll get an accurate enough estimation of how much it would cost in a particular reason on Azure. So it's a good tool for just local development and testing and proof of concepts. We will call it blog. And then what we'll do is give you the new containers. So just click new container and we'll choose the container ID. Container would be posts. And good partition key, partitioning key here presents would index value to use. That will be blog ID. So we'll just click Okay. And then if we drop down posts and looking at items that you see here, so you'll have the ID of the post and the blog ID is the partition key. While we're here, what we'll do is put in a blog post and you can put in a few blog posts, x2i contest bullet here. I'm just going to put in an ID, blog ID, both being one. The title, the content comments block, just something to make it look like. What are real blog record our document would look like in Cosmos DB. You can just go ahead and do that. You can just click new item. You'll get this editable section and then you click Save when you're done. And then at the end of the operation, your record will get some more metadata. You don't have to worry about that. But that's what the record should look like in Cosmos DB. Lets us go over to our function dot JSON file and add the configuration. Now there are two ways to add configurations, and I'm going to do it the manual way, or at least show you the manual way because based on the tooling and based on the project template that you may have spooled up, you may or may not have this option. For instance, I'm using the TypeScript functions, project or template. When I right-click, I see the option there to add binding. And what this does is it brings up little wizard that allows me to fill in the different points of configurations based on the type of configuration I'm doing. Not all of the project types have that because I've noticed that it's not there in a C-sharp functions project or attempted. So I'm just going to show you what it would look like. Generally speaking, whether you're using TypeScript, Python, Java, C-sharp, etc. In the form of JSON, this is where all the bindings are configured. So we have here, we have one for the N. So when you see it, it means it's an input binding. When you see means that's an output binding. So we already had two by default, nowhere adding another one. So you can just go ahead and add the comma, starting up a new object block. And then the type is Cosmos DB direction is in the name is input document. You could change that to, you could call it blogs, whatever it is, what this name is going to play a role in the code itself. So I'm calling it the input document. If you were using the default, the wizard would have defaulted to the word or the name input document also. However, the database name is blog. We created a collection called posts. So database name, collection name. Then we have connection string sitting. This is going to have the same name as the clinic shown stream that we had created earlier. And then we have the ID and partition key node ID. Here is going to refer to what binding parameters should I use to get the ID values. So when I see query, this is like a super global at this point. To see this query is going to be akin to the query string that's coming in. So remember that the default template for an HTTP triggered function would have the request dot query dot the variable name. Or in C-Sharp it we'd be looking at query. So pretty much query is just what it recognizes as a query string variable. So a query string look for a variable called ID for the partition key, we're also using ID stories and the same thing, no, they are both in curly braces because it's more like a binding. Binding parameter as opposed to value. Those are the binding parameters we want. No. You could actually specify like an SQL query here as one of the key values SQL query that would actually allow you in the context of a Cosmos DB and potentially other database related input variables, input binding, sorry, you could actually write your SQL statement here to get them more complicated SQL statement if you wish. But in this case we're keeping it simple. All we want is to pass in the ID. It should go and fetch it from Cosmos DB for us, and then we can process it afterwards. Now I'll just give you a quick demo of what the wizard would look like. So you just right-click function on the list. Say at binding, you choose, you choose the type. Let me zoom in so you can see better. And then like I said, we're using Cosmos DB, you would choose the name, so I'll leave it as the default. You choose the name of the database, you choose the name of the collection. You choose from the connect, the connection string from your Settings, and then you give it the document ID. So that would've been this binding. So it'll said query. Just to see him weird the partition key, I would've said, Where am I getting the partition key value from? Because I'm using the ID for both for the partition key. That's why I say query dot ID. However, if it was a case where I should pass in the ID separate from a partition key. Then I could say key or whatever it is. Whatever is the name of the parameter that I'm expecting as a partition key I would specify here. And then pretty much you just go through and then I will press Escape to cancel if I don't want it to go ahead and proceed. But once I would've press in dread that SQL query would've generated that block. Regardless. It would also have put in the SQL query and it would have been blank. And I would advise that you just remove that because I've seen it causes problems. The problems aren't really very clear. If you used the wizard, you would've gotten an SQL query. And it would have been blank. My advice would be if you're not winds of using just remove it, you can proceed with less concern. So now that we've configured or Cosmos DB, we've configured the binding, we've configured our connection string, let us write some code, refactored this method for you. And I went to point out all the changes. I mean full, you can just remove everything between the first contexts log and everything up until the last, the closing brace, right, so just empty or the function. So it looks like this. Sorry. Should look like this. When you finish emptying it, you can leave that if you wish. No problem. First modification would be to our function parameters. So we have the contexts, we have the request. I know I'm adding the input documents. So this input document, and because I'm using TypeScript, it's very strongly typed. But it does allow you to be able to lose someone to see any here. But we could specify a data type if we needed to, but I'll just leave it as any context requests. I know I'm adding input colon, sorry. And then whatever code was there from before. No, inside of the body of our function autumn going to do is say if not input document, then say that we couldn't find that ID, then we can log that message, but then I'll be responding. So contexts dot response or risks, I'm assuming that means response is equal to. And then we can specify a status modifier, the status. And even in testing in earlier parts of this course scene where we pass it in a wrong value and we still got a 200. Okay, even though it didn't find the value that was supposed to be there, which is wrong. Http responses need to be very specific. So if you enable API development, you will be able to appreciate that. If not, then just know they need to be specific. So even my saying 5400 here is wrong. And thus more like a 404404 means not found. After all, I am seeing it wasn't phoned. 500 would be if it was like a fatal error on my side, which is not the case in this particular scenario. If not input document and nothing was provided here. So this is automatically sitting down and watching for a binding value coming in through the request. That much is the binding that as specified here. If no binding Valley came through, then if that binding value didn't find anything in the database, because what happens is once it sees it's automatically going to go to Cosmos DB and check against whatever it needs to check again. So if you didn't find anything pretty much at a time, so by the time it gets here, it will know if you've found anything or not. It's in JavaScript because it's basically saying if it's not, then we say not phoned, and then we respond with a 404 and the message, it wasn't found. This in the next line. We're going to be seeing respond with a 200 and the body should contain whatever contents are in that input document. All right, so remember that's just the JSON block of JSON. So let us see what that looks like. So I'm going to run, as usual, I'm going to use Postman to run that test. So I'm passing in that parameter, that query string, ID equals one, and I'm getting that 200 response with the content of that document ID that was not present. Let's see. I passed in ten. Then I get the four hundred four hundred phone down the message blog post and not found as many documents as you might have added, which with whatever ID's, etc, it will just automatically go in. So that's when we see an input binding. As soon as the method is executed, I'm supposed to bind to this particular value. And it is for the purpose of querying this particular resource or interacting with this particular service. In this example, or input binding is looking for an ID as per our configuration here. And once it sees that ID, or once it gets called, it's going to triangle to the Cosmos DB to get the matching document. All right, let's continue along this line of thought and add an output bindings. So I've already added the configuration and you can just go ahead and pop that in there. So a new object block. And then the type this time is Q, the direction is Alt. We have the name. That's the default name that would've come up in the wizard, the queue name, we give it a name and then the connection, and this time it needs storage connection. And we have the Azure Web Storage, Web Jobs storage. A connection there? No, for context for what all of that is pointing to. If you bring up that Azure Storage Explorer, you're going to see here under local and attached to a storage accounts. So a storage account is pretty much where your blobs would go. And then it also has queues and tables. So we're using the cues that will be formed in a storage account. Of course, we're emulating it right now. If you went and provision on this on Azure, then you will be able to access it down below here. However, locally we have it emulated so we can go to the cues and the old Q, so you don't have to create this Q when you execute the commands that we're able to write. If it doesn't already exist, it will actually go ahead and create it for you. That's the beauty of it. So you don't have to create it right now. I'm just showing you where to look out for the creation of the old Q. If you were to go through the wizard for that binding the let me just show you what it would look like quickly. So let's say that we wanted an alt binding Wu tools that we want Queue storage. Then they would ask us for the name. So there will be output queue item to whatever the name is. They're probably didn't give it a different name or it really doesn't matter because that will just create a new queue section anyway. But let's give it a different name. And then we would want to point to the storage connection string, which because we're doing local, we're seeing use the development settings. And then you'd be able to proceed. And then that entire configuration block would have been generated for you. Once again, you have two ways to do it. If the wizard is not an option, then you can always just get familiar with those configuration. Keys. In our quarter are going to make one adjustment. We're inside of the section where we don't find our blog posts. We're just going to send over a message to the queue. It's just going to send that response message over to the queue. So I went to see a context, dots bindings. Remember contexts dot bindings is going to say contexts. Go and look for all of the bindings that you have. And then get me the output queue item. That's what we called it here. Then the response message is what I want to put on that queue or whatever misses. So you could put a different message. It's a messaging queue, so it's generally text-based. That's really the only change that we need to make when we run and try to find a blog post that doesn't exist. And so far I've been testing with ten. We would see that we do get back the response just the same. But if I jump over to the Storage Explorer and look in that Alt Q, I'm going to see the message or messages being deposited on the queue. That's pretty much hole. And alt, Alt binding, output binding would work. So the input binding basically stone on watches and says when I get this value, I am automatically going to fire a trigger to retrieve something or do something. And output bindings is, this is what I'm going to send out at this point in the execution. So like we saw that you have a number of binding options. You can do an SMS, you can do a signal, our Update, Hub Update. You can do a sin grid, which is an emailing service. And you can do HTTP, which we have been doing. That is an issue, GTP binding protein months. So contexts dot raise is basically saying in context, go and get the rays. That's the name of it is short for response. I guess if you wanted to rename that, that's all that would take. So that is really it for input and output bindings. We need two very simple examples. And in doing these examples, we also looked at different third party tools that we can help use the help in our interactions. So even in Visual Studio Code, I went all the way over to the emulator for the Cosmos DB. But once you have that installed, you don't necessarily have to go into the the actual management screening could actually do it from right here under the database is section in Visual Studio Code because you would have seen add our attach an emulator here. I touched ME lethal. It would ask you if it's core or DB for Mongo API you select for and it would automatically coding the local Cosmos DB. So that's a nice, easy way to a 100 that operation take and create a database and the collection from right here. We can also do virtual machine management, but we don't have any VMs in this course. And you can manage your functions from here, etc. So there are tools. Once you understand how they all work together, then you should be fine. If you're having any issues with your storage account, which you might one, I would just suggest that you update your Storage Explorer, makes sure that it is the latest version on retinal. I'm running 1 to two. I went I was on 1.170 some issues because they've made some updates since 1 to two. And you should have the azurite milliliter for local Azure Storage development installs nobody time you're doing this recording, you probably won't have these issues. But if you are working with an older toolset toolset and then you're not installing fresher than I would advise you to just install fresh because that caught me out. One, the storage emulator is what would've come with the original version of the Storage Explorer, but that hasn't been deprecated since azurite is the new thing and it should've come with Visual Studio 2022. However, for me, it didn't quite work out that way. For whatever reason, I instead use the npm to install. So you will have already had npm or Node JS it version later installed. If not, you can go ahead and get it. But in order to keep up with the activities, if you were using TypeScript, just know you should have had it installed already. You can just bring up your command prompt or your console and run that installer that it installed in the background. And then you'd want to create a folder in that location or wherever, and then run this command so you can make sure it's running in the background. This will override the previously installed Azure Storage Explorer emulator. I'm only showing you these things in case you're encountering issues because issues are natural, but it's always good to have the solutions. You can do all of that and then restart your storage Explorer and then everything should be able to connect. The symptoms of it not working would be that the function will not execute. What you try to test. It will pause. And another symptom is that when you try to connect to the emulator, it would tell you about connections being timed out our refused. So if you're getting those symptoms, then you definitely need to go ahead and meet these updates. That being said, that's really it for input and output bindings. So I encourage you to explore, creates other functions that you think would be useful. Timer triggered function that would send off an e-mail every morning or every week, stuff like that. So go ahead and explore. 21. Durable Functions and Orchestration Patterns: Welcome back guys. In this lesson we are going to be talking about durable orchestration or deal with functions. So that's another function type that is available towards in a 0 functions. We can easily spin it up using the template and we will be looking at it. But before we get there, I just want to go through it. A write-up on its on the Microsoft documentation for Durable Functions, durable form of stones, That's an extension of Azure Functions. And it basically uses the orchestrator pattern to follow through on things. So just for context, I durable form shown is one that you can use the define workflows using procedural code. So our flow generally is you want to do this then that, then that then that you may require Monday while the intervention in between R-naught. But the point is that you don't want to get to step three unless steps 12 have been completed successfully. So a durable function will actually truck the state of the operation so that it knows what decision to make with each point of reference. They are also capable of calling other functions or other durable function synchronously and asynchronously. And the output is reliably stored in local variables. Our history as a functions are durable and reliable and they basically checkpoint each time they do an function or weight or yield. So each time a function is called and it returns its value, it creates such a point so it knows exactly where it is in the operation and the state is never lost. It's always there. It's always keeping track of what is happening and where it is in the process. They can be long-running, they can be seconds, days, months never ending. So you have different patterns where you might be calling an API asynchronously over and over. You might be pulling a data source for something and you wanted to continuously do that. The most important parts of this is wanted trucks the state. And it always allows you to know where you are in the operation. If you want a better understanding, at least with the theory, you can read through the document is and more. And you can see some of the code examples in the different languages. But we're going to be doing it together and we'll be using Visual Studio Code and a C-Sharp template. So of course we jump over to Visual Studio Code. I've already created a folder, a 0 durable functions. That's what I called it. No, we jumped onto the Azure tab and onto the foam Oceans section. We're going to be creating a new function up. So we get to our usual wizard. Me, just zoom in so we can see clearly, I'm going to choose my folder and then I'm choosing my language. Then I want our own time. So I have six installed. Of course you choose runtime according to your context and of course the version would be relative to that version four goals with dotnet six. And then I want you are able to form shunts the findings easily. You can always just start typing and it will filter it out. So I wanted to durable functions, orchestration. And I'll leave the default name for no and default namespace. Then they wanted to know which storage account I'm going to use the local emulator and allow it to generate a project with it created and we have our new code file. I'm just going to zoom out a bit so we already are familiar with what the project looks like. If I look in the CS project file, you see here that I have this particular library which is very important for our durable operations. So if you were using Visual Studio, just double-check and make sure that this package is being referenced. If not, you can go and fetch it on new gates before you create the durable function template. That's fine. But then when we have our code file, weekend, just look through and we'll just look through and do a test. I'm not going to modify it too much or anything. We just wanted to understand what exactly is happening here. So we have the annotation that tells us or functional name. This function name has, sorry, this function has a function called Ron orchestrator which is returning list of string. And then it is passing in an orchestration trigger. And you can see here the context is of type I caught I durable orchestration context. So it's tracking all the outputs. And then it's going to say each output, the value that is gleaned from calling this function with that value passed in. So you see here, it's only one function that it really has and that is the durable functions, orchestration, C-sharp One on the score. Hello. It's only one function which is defined here. But imagine if we add multiple functions that you wanted to call one after the other, you could just define them in that specific order. So here it's saying call the function by that name and give it that value. This value is going to bind. Here's our function that is calling. And even though the name is here, the function definition is called sayHello. And then it takes an activity triggers. So this activity trigger is pretty much that parameter that is going to get boned. You can woodwork like an input binding. It could be something for our Cosmos DB or something else, you know, whatever it is. We've gone through our aisle bindings. So we have an appreciation of how that could work. But once you do that, once you pass in that value, then you carry out the operation based on the value that came in and do whatever. So here it's only going to say hello to each of these languages, each of these cities, sorry. It's just going to return hello with that volume. That's all it's really doing. No. The last function that you get here is one that is triggered through an HTTP trigger. So we know about HTTP triggered functions already. So that means when we call it, when we call this Durable Function and this whole durable operation through that HTTP trigger, I noticed your book client is calling I durable orchestration client and you call it Starter. Then it's going to start the operation on stored at instance ID. And all this is doing is in starters, start a new async of the durable function, which was what we looked at first. Http trigger is going to trigger the durable function. And the durable function has its operations outlined, trigger this. That many phoneticians in that order stored their outputs and make one big return. The end of all of that, it's going to log and see it started with the instance ID. So the incidence ID once again hopes it's a truck that this operation is currently in progress or it's finished or whatever. And you can use that in your logs for analytics. And then at the end of this whole triggers just going to see check startled response and give a status, the instance ID. So the incidence ID once again would be the incidence off this book are castration happening? And then it's checking the status response to see is it finished? Are we okay, whatever. So like I said, it stores the state of me explaining, let us see this in action. So I'm going to run this function and I went to debug anyway. All right, it's on the ones that's running. You're going to see here in the logs that it is detecting that it has the HTTP starts so that these are HTTP triggered method. Then we have the arc iteration, and then we have the activity. So that's what we'll call the function that it calls. It's an activity function. And this is the art history. It's a film. This is a simple HTTP triggered foam shown. So when I make that call, HTTP starts and I'm willing to control click, Control clicking of course would have launched it in the browser. So my browser launch, he called URL and we know what happens when you call it the URL for an HTTP triggered function. And it actually went ahead and start the darkest tuition, right? So this one it was called, but the host APIs. And then you see here where it says hello was called successfully and it returned hello Tokyo, hello Seattle, and hello London. That's evidence that it was called successfully. But look at this also. Let me see if I can zoom in a bit so we can see those logs better. You'd notice that each time it would have successfully made that call, it goes right back to that art history OR function. It calls, it's executing the activity. Then we get the result of the activity. Then it tells us it's executed the activity. And then it goes right back to the Arcbest treat or every time it makes a phone call, it returns control to the orchestrator. And then of course, you are the one working on the workflow. So you put in your logic and your logging and your decisions based on the output or the success or lack off their lack thereof. Off the durable function, calling the activity. That is hole. You can use that to your advantage. When you called HTTP triggered function r dot URL, you'll notice is that you get this block of JSON. So do you use postman or used the browser? It doesn't really matter, but what you're getting is a status query gets URI value, which has a URI that will give us the state, the current state of the art history. All of that. Is that URI I went to copy that and let me just place it in a new tab. And you see here it says name, we get the name of it, we get the instance ID. For sure. We get the runtime stack. So switching this case is completed. So at any point in the middle of it, if you want to see what is the status of this durable function, Is it done yet? This is still in progress, then you would be able to just see, give me that status. That's pretty much it for durable functions. Of course, it's a really simple example, but I hope you're seeing the possibilities and how you could actually use this kind of workflow for more complex operations where you need to truck. Is it done yet? What is the status in the middle of it? And make decisions based on all come off one operation going into another. 22. Create Serverless REST API using Azure Functions: Hey guys, In this section we're going to get down to some fun stuff. So we've been looking at some lots of theory, a lot of concepts and just getting to understand the inner workings of a 0 functions. In this section, however, we're going to put all of what we've learned, the test and we'll be building out a rest API. So we'll just be using functions, building old endpoints. If you are not familiar with rest APIs, you can trickle my other courses on rest API development where we actually go through how to build them using a whole EPA development project template from Visual Studio boat here we're going to be using Azure Functions. In this video. Let's just get started. We're going to be using Visual Studio for this one. Let's just find the ASOR foam zones. Once again, if you don't have it to the left, you can always search. And if you don't have it at all, you will have to go and get the installer and install the Azure workload. Let's get Azure Functions. And this one I'm going to call, I'm going to call it a shopping cart list. I think that's a simple enough example. We have items that Linda shopping cart and we take them off when we have them. We can just go ahead and create that. And for now, I'm going to do an empty template for the storage emulator. Yes, we wanted to use a storage emulator. We already went. That works. And just for context, this point, the zeros, traditional Azure Storage, Emily told me and not necessarily to work properly. And I would've mentioned that you would have to use azi, right. So to make sure after you have installed it that it's running, you can always just run that command azurite and then you'll see it starting up all of the services. So if you encounter any difficulty connecting to your trend to use these emulated services and sending it can't connect, just run that command azi, right? All right. I'll just have that running in the background. Foreign. Oh, let me go ahead and create this empty template. When we come back, we're going to get started with creating or functions and looking at the concept of hallway are going to lay it out in terms of building out our rest API like structure. 23. Setup API Routes: All right, so let's get this started. I'm going to add function, so new Azure function. And since we're dealing with an API like structure, then all of my functions need to be HTTP triggered. I'm going to call this shopping cart function or shopping carts API. Let's call it that and add. Then I'm going to go with the HTTP trigger. I went to make it anonymous at least for now. And click Add. All right, so nowhere more functional, not when to start modifying just yet. Instead, I'm going to go ahead and put in some other things. So I'm going to add a new folder, I'm going to call it models. The thing with API development is that you typically won't. Models are details that govern watts data you expose slash interrupt with for each request. I'm going to have a shopping cart model, first of all. And I went to meet this internal, internal means that it can be used by anything inside of the same project so I can leave it as internal or no. And I'm going to add some properties. I just went ahead and did all three clauses and you can just look at them and replicate them accordingly. So we have the shopping cart item or a shopping cart in general. Actually this should be a shopping cart item. Let me just update the name of the file accordingly. Let me just do the shopping cart item has an ID which I'm just going to use as a string GUID. I have a deed created, and then I have the item name, and then I've collected or not, you know what, like on our shopping list, strike off the item once you've put it into the cart. That's pretty much what that is. Signifying. Then we have one for decree. Of course, for the Create. I'm not willing to expose everything. I don't want the user to be in charge of the ID or to create it. And Booleans default, default. So we don't miss specifying that are needing that from the user. I'm assuming that it's not yet collected. All we need is the name of the item. For the update. You may be changing the name of the item. You might also just be changing collected or I could just upload them to just see yes, it's collected or not. So I think I'm going to keep it that simple and just work with that later on when we're connecting to the database, then we will have to create some more models. But for node this is all that we need. A rest API. These would be what you call details or data transfer objects. Once again, if those terms are familiar with them, you can always check out my course on API development. However, it has done back over to our function. Then what we're going to do here is actually establish or roots. So the thing with functions is that by default we're getting this class, yes, and then we are getting one function with R1. In this case it's saying you can do, you can do post. We don't really need anything to do both because in our rest API sitting, it should be very strict. One for posting one forgets already saw at the very basic level. But each function should be in charge of one operation at that time. So since we're mimicking that rest API structure and standard, we went off to modify this. What I can do is rename the function. So if this is supposed to get shopping cart items, I can call this get shopping cart items. All right, so that's the function name. However, the route, I want to specify a value because I don't want not when you call the function, you have to say slash api slash this function name and then that function in once again rest API standards would have you call this one EPI, one structure. But based on your verb on potential parameters, values being possible over then it wouldn't know which function exactly you're referring to. So we're going to rename this route so that it looks a little bit more like what our rest API would look like. I'm going to call it a shopping cart items. So your expected because it's already going to be the URL slash, api slash and then the function name. But when I specify this as a root, then it wouldn't know that from the HTTP request, anything going into that root after the slash API will be going to this function, right? Next thing is that I don't want this function to post. I wanted to retrieve some went to remove that post. And just to make sure everything looks good, I'm also going to rename the function itself to reflect what that function name is so that everything looks kind of uniform. Um, for null, I'm going to remove all of this. I can just remove all of this from the function because what we're going to be doing require all of that per c. That is the one to get. No. Once again, we're doing our RISD API, so that means I need one to create. So I'm just going to copy all of that and I went to call this one create shopping guard I attempts, we need to give the method that name. And this one is going to be the post because it's not getting this time. It's creating an entire me create. You should use post and then it has the same root. So it's going to be a post request to that EPA or that HTTP route. This one is gets requested that HTTP roots. We can do the same thing for PUT and delete. Put would be for updating, deleting. Nothing that's self-explanatory. Food shopping cart I attend. Or let me just make sure that my methods are really reflective of what the operation is going to be. So delete shopping cart, I, uh, Tim, delete shopping cart item. This one is a PUT request and this one is a delete request. All right, Once again, we're just mapping old roots. They may change as we go along. We might see something that we wanted to change. Of course, we're going to be changing the parameters when we're focusing on each one. But for me, I'm just mapping them out so we get the big picture. There is another type of gets because maybe we want specific ones. So if we wanted a specific shopping cart item, then it will be get shopping cart item, not a tens. This would also be a gets put in for the root would be shopping cart item slash. And then we're going to use a binding parameter to let you know that they should expect something called ID. Afterwards. I'll get all the items we have. Gets by ID for too much yet shopping cart item if you want to qualify, just say you're very clear and you know that that is what it's for. But all it is, That's the ID. We have the Create, which is just a post. And then we have the PUT and delete node that we're good idea of the big picture. When we come back, we can start working on them one by one. 24. Setup POST (Create) HTTP Azure Function: All right, so we're back and we're going to start off with the create shopping cart items. The first one, let's update or log message. Just going to say creating shopping cart item. In the request, we do have the data coming in. We've already seen that this request Good of looked in the body, in the URL. And since it's a post, we're expecting the data to come in the body. I'm going to just start off with a variable called string requests data, where we're going to await new instance of the stream reader. And this stream reader is going to go into or HTTP requests parameter req dot body. Then we want to read, read to end rather async. Alright? Now that we have read it, this is going to return that JSON body. Pretty much we want to serialize it into our shopping cart item object. So I'm going to save var item is equal to, I'll use JSON converts. Json convert comes from our Newton's soft library and that was already included. Here. It's basically a requirement for HTTP triggered functions. We've seen it, we've used it before. Well, my point is that it's like any other pocket is just a NuGet package. So if you already know how to get packages and you can just go to Manage NuGet package and add it right here without any qualms. We want GSM convert dot de-serialize object. We wanted to de-serialize into our create shopping cart item object. And then we give it the request data, which we know is JSON in, in nature. Then after we've done all of that, we would want to edit some some database. Our database is going to be of type shopping cart item because we want the defaults. Before I do all of that, let me just create a new instance. Okay, let me change this from item to data. And then I wanted to say var item here is equal to a new instance of the shopping cart item. Shopping cart. I turn. This new shopping cart item is just going to take that item name, which is coming in from data, Data dot item name. Once we have that object that we would want to see if we wanted to save it to a database. I'm not yet ready to start the whole connect to database stuff. So what I'm going to do is just use a static list, but I'm calling a shopping cart items is equal to a new instance. So down here I'll just say a shopping cart, items dot Add, New Item. Now, let us pause for a minute and reflect on the relevance of using this list. What the pros and cons are obviously in a real setting, wouldn't be using a list like this, would be using a database. However, I'm not ready for the database just yet. So we're going to be using the list. Now the thing with function up is a While it is running, it is going to be storing any data, any variables during the runtime. So while we're testing this is perfect because we can call all the different methods and would be interacting with the same list. A production setting. Remember that functions are able to scale. That means that if you were to deploy this while it may work for some scenarios while your function app is running. Because remember that on some plants it will actually be turned off until you have to split it up again. So that's one danger, but to let us assume that doesn't happen, it can scale to have multiple instances. So then you would end up with multiple instances of the list, which is another danger. So of course that's why we would want one database support for null. We'll just use our list. After we added to the list, we can then return a new object results. And I'm just went to pass in the new item, the whole list, but just the item that was no created. That's really it. All of our errors for our create shopping cart item are gone. And that is step one. So when we come back, we'll be going into our get's. We'll just do both gets in one goal, gets all vs. get by ID. 25. Setup GET HTTP Azure Functions: Alright guys, so we're back in this lesson, we're going to be focusing on or get methods. The first method, which is to get all items that one is fairly straightforward. So my message is going to say get, getting all shopping cart items. Then all we're really going to have to do is return. So at this point would really go and query the database and say give me, all right, In this case we don't really have a database, we only have this list. So all I'm going to do is return on, okay, Object result with the list. That's the list of the shopping cart items. And once you want, all, you get to all. So that's pretty much it. I have an error here. Sorry, I missed off my new return, new okay, objects results. There we go. That's really it. For the get's by ID. It's a bit, it takes a bit more finished, right? So yes, we have the same rule. We know that this one is expecting an ID value coming in. So that ID value usually comes in as a parameter. In an HTTP request. It could've come in as the HTTP or parameter. I'm going to add a new parameter here as a value stream ID. Then I'll just use the comma string ID is going to represent the ID value that should be passed in when this function call is made. I'm actually going to put this as the last thing. I'll put it after the logger stream ID. Once this method is called, we're expecting this ID. Now once we have that ID, and I'm just going to rewrite the logger message, getting shopping cart item with ID. And then I'm just going to encapsulate the id value that's being passed in. Sorry, in interpolate, to encapsulate, know that I have that. I'm going to say var shopping cart item is equal to, I'm going to go to my shopping cart items list. Then I'm going to see dots we could use find. But I'm going to stick to what I am more comfortable with first or default because I want to do first or default. If it doesn't find it, it will return with a null. First star default and then Control dots to use system.in link. And then we put in our Lambda expression. So q dot ID off the shopping cart, shopping cart item will be equal to the id value that is passed in. And then once we have done that, I went to know check. I see Visual Studio is helping me along. If the shopping cart item is null, meaning while obviously it didn't return anything, then we would have to return a new not phoned object results. So that's a 404. Otherwise. Otherwise. Okay, so here's a difference between, because you might notice that you have object results versus results. Let's see of old key results different from Old Key Object results. The difference is that when you return to object results, you need to put in data when you're not returning detail. So in this case I have no data to return to its expecting some value. I have to say not phoned results. So they really wanted to see him thinking it's really the same four over four, but just which one to use when appropriate. We are going to return a not found result. Otherwise we're going to return a new object results with that particular shopping cart item in hand so that the color would've gotten the detail. Let's review quickly to get multiple items, get shopping cart I attempts, All I'm doing is running out query quote unquote, running a query against the data store and returning all know what a datastore is at least. So I'm just returning what is in that list. When we want by ID, I have to specify that I want the parameter ID to be prisoned. They're already both get, we already know the roots what this one is, slash IID, hence my need for the ID. And then once I have that idea churn rondo query to find that one object in the Datastore. If I don't find it, I return not following results. Otherwise, I return the old key object results. That's really it for the get methods off our rest API. 26. Setup PUT (Update) HTTP Azure Function: All right guys. So let us jump over to our PUT we're looking at updating or shopping cart item. No. We have to make some more changes because when we want to update, we would have to pass in an ID as well as the data to be updated, right? So we went out to do what we did up here, which is include ID, ID parameter. There we go. And we have to update our roots because we also need that stuff ID on the end of the update roots. Then we can GnG bar log message, updating shopping cart item with ID. That's in. No, we have to do kind of a hybrid operation. The first thing that we have to find the data so we can repeat what we did in the Git. I'll just go ahead and see getting me this shopping cart item that has the id that was passed in as a parameter. If we didn't find it, then return not phoned. Alright. Next we are going to have to repeat something like what we did with the create. Some went off to parse through the incoming data. This would be parsing into the update shopping cart item. And then in a bigger operation, we could use something like ultimate bird to map between the incoming data and the shopping cart item to do the updates. But this is a very simple operation. So what I'm going to do is simply update the shopping cart item, dots collected flag, and let u be equal to the data dot collected flux. So you're updating it. It must mean that you're changing the Boolean from true to false or vice versa. That's all we're really going to be updating. Then at the end of all of that, we're just going to return okay, object results. And I will just return the updated shopping cart item. The object that's really all there is to the update. So let's go through it again because I kind of took code from different pieces. So let me just explain what we're doing. One, we're finding the shopping cart item with the matching ID that was passed in via the parameter. Then we're going to say if nothing was found, we return a 404. Otherwise it will continue to see get me the data that came in in the request because with a put, the data would be in the body. It's very similar to how it would be in the body for the post. Give me the data from the body, piracy tilt into the corresponding model for updating. And then we're going to manually update. So I'm doing it manually in a bigger operation, like I said, you would probably use automatic or, or some other third party library or to convert or to have that done for two very simple operations are very simple operations, so we'll just do that. And then we return the updated object as part of the return results. That's really it for setting up the PUT. When we come back, we will do our final one, which is the delete. 27. Setup DELETE HTTP Azure Function: All right, so let's get started with the delete operation. So I think we've been doing this long enough. If you want to hit Pause and try and do it yourself, then I would encourage that. Otherwise we can just walk through it. I already wrote the code. We had sets up the delete shopping cart item method from before. What we didn't do was official to root. So the root once again will require an ID value there. And we have the matching parameter string ID updated the log message. And once again, we tried to find it. If it is not there, then we say not found results. However, if we continue, we say remove it from the shopping cart list and then return an old key results. Note not okay, object results, both Okay, result since there's no data to return. And that's really it for the delete method, right? So with that, if we just do Control Shift and B, we can build and make sure that we have no errors anywhere in our code. We can see that we have a successful build 0 errors and we can continue. So now that we have fleshed out all four off our operations, when we come back, we'll just go through some testing. We use Postman to call the different endpoints and make sure that it is functioning or that our functions are functioning the way we expect. 28. Test Azure Functions with PostMan: Alright, so it's testing time. We've done all the hard work. Let us put it to the test so we can just click the start button here in Visual Studio to start to talk. Or if you're using Visual Studio Code by now, you know how to run that. And then in our console, that accompanies the execution, we get a list of all the functions, so we see all of them by name, thanks to the function name attribute that we added. But we also notice that each one has a root. Yes, the rules would have been based on the name that was provided by default. However, we overrode that and we'll put in our own roots. And there you can see how you can call each method according to the route. If we test with postman or any other HTTP testing tool, and I say, okay, call the get all methods. Notice to get all shopping cart items, I will just call that URL so you can actually just highlight it and then copy and paste it into Postman click Execute. And then you'd see here that you would get that stuff, those 200, okay. But nothing because we haven't put any data in. Of course, you'd also see the log gets updated. So getting all shopping cart I attempts. And then it was executed. At that point in time. Let us try and create a new shopping cart items. So I would just change this method to post. And then in the body after make sure that I send over some JSON, so changes up to RA and then James F to G is on. And what did we specify? In order to create? We are only accepting the item name. So I'm going to see item name then colon. And remember to put in quotation marks, then colon, and then this would be just from Postman. So when I click Send, Let's see. Then you'd see it replies with the newly created IID, that timestamp, the item name, and if it's collected, that's false. If I retry my original ghetto, just open a new tab, uses the same your bullets were using a get and I send. Then I'm seeing that I'm getting buck that one item from the list. If I try my other get, This one requires the ID. So I went to see a slash and then passing that ID value, copy, paste it there, then I should only get back one object. There we go. I think that this is working well, let me try the update. So I'm just creating tubs accordingly. So once again, we have this URL, paste it there, or actually I should have just used this URL goes for the updates. We do need that ID value, going to do a put. And then inside of the request body raw JSON, I'm going to see that collected is no true colon and the value true. Let me send, you see here it returns, they'll be heated object Latina know that it is no true. So if I go back to the gift, gets all send again, then I'm getting back the updated. Like I said, if you call these functions within the same session, I went to call it a session. So we're executing, right? No, we're calling one after the other, one after the other. It's updating the same list each time. So that is why you're getting those results. Of course, if we stop and restart, then that list is going to be emptied out. So it's the same way That's a foam shown up once it's running. If you're using that list, then no problem. Once IT skills are winds down and then as the wind backup afterwards, you just know that you're going to lose that data. So you don't have to be careful with that. But later on, once again, we're going to be looking at using persistent storage in the form of a 0 tables are any other storage medium that you prefer. So that is really it. So the last one that I'm going to do is I delete. So I'll use this URL over here. And this time the method type is delete. We have the ID and everything. I don't need a subject body responds to 100 old kid and all content. If I go back and try to run a query and nothing comes back, if I recreate, then I get a new ID. And then if I tried to look in there again, I see everything. That is literally all there is, quote unquote all there is to our rest API. The more difficult it becomes this Bs based on your business needs bullets at the end of the day, this is how easy it is to create a whole rest API using just Azure Functions. 29. Test Functions with User Interface: Now a fundamental part of any API is development, is how it interacts with client apps. That can be a mobile app, it can be a web up, etc. In this case, I made a simple blades are up where ruled code to interact with the different endpoints from our z are functions api. While running it. We know that we have all the roots are already. All I did was retrofit the Fitch data page from the typical blazer up template. And what it does is test from laser when they put you in an item here and say Create, it will actually update sometimes what happens is that when you interrupt with the console here, it will cause it to positive. You just press enter, it will continuous. You see here it actually made the call create shopping cart item. And then it needs to retrieval. And there's our new shopping cart item. If I say a test from blazer one, and let me just make sure that the console knows that he should operate normally when I click Create updates and nuts low in the list. If I click the tick for collected, you see here that it will call the port. Here we go. So put it, updated it. And then if I remove, once again, I keep on interacting with its pauses. When I click Remove, it should remove the item. Let me try that again. There we go. So let me create a few more and then let me delete them. Each time I click Delete, you'll see it updating the interface. So I'm just going to walk you through the code that I wrote for this. It's not really a blazer course. I'm not teaching ultimately the bizarre up of other courses for that. But just within the context, I'm going to show you that from the blazer standpoint, it doesn't know or care what the underlying infrastructure of this rest API is. It's just going to be doing, is there stuff relative to our rest API? I created a brand new project and it was just a pleaser Web Assembly up. Simple ways to do that. You just go to the solution, right-click Add New Project. You look for blazer WebAssembly up. I'm using dotnet seeks go ahead and create it. And then by default you're going to get some address here, BCE address with environmentalist that she is told to be the base address of where I know our functions up will be broadcasting from. All right, so we know that that is the port. We would've done some testing with Postman. We know that that is our standard API wrote, alright? So that is the base address for the HTTP client being used in our beliefs are up. Another thing that I wanted to point out is that we had to set up cores. Cores is cross origin reference. Okay, I'm sorry. Here's the definition. So Cross-Origin Resource Sharing. There we go. That's protocol or sets of rules that govern your API interacts with resources that are not on the same server. While we were running the function app anticipating with post-money didn't see our problem. However, when I'm broadcasting, are sending a request from an external source that these blades are up when it's executing, it's executing on a different port. Therefore, it's seen as a different app altogether, then it would actually reject it. So this is all we enable CORS on an Azure Functions up in our local dot settings JSON file. We go to host and we open up a new object block where we said cores to be star, meaning accept all requests. They don't have any restrictions. Of course, when we deploy it to Azure, that will be handled on the Azure side. But for a local settings, we have to make sure that is in place. You can go ahead and do that. I also would have replicated the models folder, so I just redid them, don't hear in the blades are up. And instead I updated the namespace to the shopping cart, at least dot blazer dot models as the name of the app is shopping cart, at least dot blazer. You will have the models or access to the modules there. Another thing that I would have done would be to update, of course, the Fitch data dot racer. Or if you wanted to make a new pH, that's up to you. But I just did everything on one page and I'm just going to walk you through of the code that was written. The page title, we have the h1 and the p tags. That's all just semantics. You don't have to do that. Then I have the edit form of top where I have that input text which I am binding to an object called item create and accepting the item name. So here in the edit form the model is Item Create and on valid submit we're going to call the method Handler create. Then I have this submit button for my edit form. Item create is often a shopping cart item. And I just have an instance of that there and handled create. Is going to just see response always the HTTP client. What does HTTP that says HTTP client that I'm injecting into my my component. So that would've been already there because I'll fetch data is actually assemble of how you can interact with HTTP client so you don't have to put that there. It should be there already. But what I'm doing in the handle create is emptying http dot post as JSON is sink, and then post the shopping cart item. And actually I just noticed that this is wrong. That should be create shopping cart item really doesn't matter in the grand scheme of things, but let's just work with it. So that should be creates shopping cart item. So I am posting a JSON body or JSON serialization of creative shopping cart item. And it's going to address api slash shopping cart items. So this API slash shopping cart item just gets appended to our base URL here in the program.cs. Then we're sending over the item create object. Then we say if it is a success scored or key or no content or whatever it is we've seen, we know the different codes represent different things. So an old key responses are successful, then we just call them a weight on initialized async, which basically just redraws the entire theme. For the uninitialized async, we just see items which is an array of shopping cart items is equal to http.get from JSON async shopping cart item in a form of RE. And we have the same root there. I'm not going to go all out with architecting and best practices in this particular example, I'm just showing you a quick win. If you want all of that you can juggle. My beliefs are courses are uninitialized. Async gets all the items and puts them inside of this object, which is designed for the items handled create basically just says post whatever data came over as JSON tool, this endpoint, and then go ahead and update once it's all successful, know the rest of it would just be me changing the weather forecasts sample code that was there and just using items are what I've said for costs and changing the table headers according to my new table headers based on the new data. Then for each item in items, remember items coming from our API call. I'm just updating or displaying, sorry, the different values. No, for the checkbox section, I have an input which is of type checkbox. I just give it a class where it looks kind of nice. The value is going to be bound to whatever item.quantity is an onchange. This is an event that means that whenever you change the value of the chatbox, I want to call this method checkbox clicked passing the item and pass in the value of the events. So because it knows it's a checkbox, that event is going to have data, was it true or false, change, etc. Where policy in that value. When that checkbox clicked method gets called, it is going to take that shopping cart item and this could easily just be, probably really should be the update. The ports, whatever object type we had used their multiple courses with the clash here because this is of type shopping cart I attempts. So I'm just showing you that I'm taking the shopping cart item object here. I'm taking the check value as my, as my event args value will have to make it an object at this point. Then I formulate my update shopping cart item object and set divided that I know needs to be set. And then I went to send over g sun async, that object type. And I'm going to be calling URL and then using interpolation, I'm just sticking on item.name DID. So that's why it was important to pass over the item. Because in doing the update to meet need data, both the original item and I can just use the ID. So I'm passing over that ID and the whole object with the data that represents the update. And then if it was successful, then you redraw. Then for the delete, we have a simple button. Give it some bootstrap class, and then onclick we have similar code event and possible with the item. So posible the details able to item to be deleted even though we really only need that ID. So we're just going to say virus bonds, http dot delete async, not as JSON because there is no JSON to be passed over it reasonable data to parse, just delete. And the URI would be APIs, that shopping cart I attempt or item.name DID and then we redraw. That is why you get that. Starts off real-time, look and feel to the UI because I keep on redrawing each time the operation was successful. That's really it. For bleeds are interacting with the API. Of course, he can get a bit more complicated based on your needs or whatever. Of course, the API could be a bit more complicated. But once again, we're just taking a quick wins. And once you understand these core concepts, you can always build on that for your future operations. That's it. You can hit pause at the different points and replicate the code. I just scroll through slowly enough that you can hit pause at the different intervals. You can see all of the code. And I would encourage you build more endpoints or more behaviors in the API and build more pages on components, interact with said API. Once again, make sure that you update your base address and make sure that you enable the course settings on the API to test them simultaneously. You don't want to go to the solution, right-click go to Properties. And then you'd want to select multiple start-up projects. You can put both the start or both, start with a debug, whichever one you prefer, based on the artistic process. That's really it for testing using a UI. 30. Interact with CosmosDB: Welcome back guys. In this lesson, we're going to be going through some of the changes needed to start talking to the database. The database that are selected for this exercise is a 0 Cosmos DB database. We already have some experience dealing with that. We looked at using it as an input trigger or input binding rather. And whole, we've just pass the data in and it will communicate with the Cosmos DB. In this, we're going to be taking a slightly different approach to just using input bindings because there are certain operations that we want to do. I think it's mixed. We can write the code three different types of waves if we use the input binding, because based on the situation, the court is going to look differently or we can be a bit more consistent. So I'm going more for consistency in this situation. And I'm going to show you guys how you can one sets up your function up for dependency injection to Hawaii, use Cosmos DB client for that total injection, which can be used all over the place anywhere else in any other function. To get started, I want us to go over to new gets. Just right-click your functions project and the goal to manage images and get this one. Microsoft built a 0 dot web jobs dot extension dot Cosmos DB. So this package actually gives you an API for Cosmos DB as well as Table Storage, but we will be focusing on Cosmos DB right now. The next thing is to create a startup class known as startup class would be similar to what you see in every dotnet Core application. So we have a blazer up here. In dotnet seeks. It looks a bit more condensed than what you're probably used to, but this is pretty much that startup class that they kind of condensed into one file. This basically bootstraps and almost services to be readily. I'm available for dependency injection anywhere in the application once the application is running R12 writing code. So that's what that does. So we created our own, I'm calling it startup.js. And I'm just going to show you the code and walk you through each line so that we have a full appreciation of what's happening. So just right-click your function up, add a new class called Startup. Not CSU might get an error or some feedback that you're creating a function type that already exists. That's fine. You can just erase all of that or anything that doesn't match what I have on my screen. You just go ahead and cart accordingly. We have the first annotation assembly, colon, and we're doing a functions startup. It's going to be type of startup. You might have to include some missing references as you go along, so you just do that accordingly. Or namespaces are shopping cart list. We know that already. And then we have, or sorry, our class, startup, which is inheriting from functions startup. Then we have privates that degreed only I configuration root configuration is equal to a new configuration builder where we're sitting the base to be the environment current directory, that's the directory of the project. We are adding our JSON file up settings.js IN true, but then of course the local settings.js on. So we know how to access that later on. And then we're adding environment variables and then we just build. Then of course it would give you an error saying you need to implement the abstract class, which is functions startup and doing that would generate this method stub for you. It's a public override void, configure. All of that is generated for you, both inside of that method, what we want to do is say builder dot services and add a singleton. A singleton pretty much means that whatever you're adding here is going to be access. One instance is going to be accessible throat your application. So one connections or Cosmos DB, the same connection to the cosmos DB will be used each time we access this dependency, pretty much right? If you want to know more about dependency injection and holdout works against a gold mine, of course is what I'm going to go through with this null. So var connection string is equal to and then it looks in the configuration. Remember configuration compile all of these elements, including our settings file. So we already have that setting as file, even though the name is different, it will it contextually know where to find what we're seeing configuration and get me the Cosmos DB connection, which is what we called it in our local settings file, right? So we looked at this already. I'm using the emulator and I'll show you in a few the database that we've created. So just make sure you add that key, Cosmos DB connection, and then we add that emulator key accordingly. So we're calling that DB connection and then we're seeing if it is null or empty, then throw an exception. Otherwise, return a new instance of the cosmos strength builder with the connection string dot build. So that is our startup class that we have configured once again so that we can inject whatever dependencies we've defined. We only define one which is for the cosmos client. Don't let me jump over to our Cosmos DB. We already have some experience with this. This is our milliliters. You can go ahead and go to your emulator and create this new one. So we have a new database I'm calling shopping cart items. Then my collection name is items. And for my key or by a partition key, I'm using this word categories. So we're going to be making some changes to the models as we have them. But remember that the partition keys like a high-speed lookup. So it kind of like indexes, your records or your documents, quote unquote, so that it can be found faster. So we're adding a new property called category, which makes sense. It's a shopping cart item list. What is the category? Is it food, is it clothing? Is it a need, is don't want, etc. So with that, we can build up our tissue on the data and speed up the lookup, speedup operations against this whole database. You can go ahead and do that. After you've created the database that those look at the changes to be made to our models. So the only change I'm making is really to the shopping cart item. And I'm making a GNS3 create. Let me go with a shopping cart item first. Shopping cart item, one of making it inherit from table entity which is courtesy of this new class, Microsoft dot windows, dot storage dot table, which comes in our new package. So you can let it inherit. You don't necessarily have to, but the benefit quote unquote off letting it inherit is that the class or the model will know actually have access to all of the properties that gets touched whenever our document is stored to see a lot of other Meta tags. You probably noticed them from when we did the blog exercise. There are a number of other Meta tags they get attached. So once it's inherited from two entity will actually be able to parse them over. No, I've also added JSON property annotation to our id req property. Sorry, this JSON property basically sees that, Yes, I know I'm capital I in C-Sharp world, both in JSON Web, please meet me a common id. The relevance of this is Cosmos DB is looking for a property with common ID. When we sit that ID value, you will actually get a bunch of arrows. You can try it yourself. I actually had a little field with Hillary. Remember that, that it needed to put this there. So you can actually try. After I show you all the code and everything tried to do are create, I need to see the errors that you get. You will complain that it's missing the ID column even though it is there. So we just see it when you're in JSON, please identify as common ID and then Cosmos DB will be able to see it and see, okay, I see the id value prisons. Of course we're not alone in Cosmos DB to generate the ID. We're doing that. So that remains the same. I've also added another property called category. Like I just said, category is going to be our partition key. And just like with the ID, we need to let it know that it is a common keys category. Of course, this just needs to match whole year rule to the partition key. So I'm just going to leave it like that. Capital C for codebook command Z for JSON. Then inside of the created, I'm also putting in that category. So when you're creating the item, you see the name and the category. And of course there are better ways to do this. You could use enums, you could use a select list, etc. But for now we're just going to keep it simple just so we can understand the exercise at time four, the updates that is actually kind of optional. So actually I'm going to take it out and I'll call it a code when we get there. So I'm going to remove that so the update doesn't change. So we're just adding category here. And for the shopping cart item, we're adding that Andy students on property as well as the JSON property above the ID. It's were coming along nice and let's jump over to our function. No, you'd notice one, I have removed all of the static keywords, public. This was public static class. I removed that because static classes can't have constructors. We need the construct of oral injection. And other thing is that sorry, I just removed that line. So if you started tapping into again disregarded, I apologize. So another thing is that we can't have static methods because we have to modify the code in there. And that means we have to use the private. Fields that are going in. As we explain what's going on here, I'm just showing you the erudite you'd get if you lift the method is static, so no static class and no static methods. Alright, let us go through line by line and see what is happening here. So firstly, I'm injecting the cosmos claims cosmos client is coming over from our start. So this is returning an instance or the dependency needed for us to access Cosmos client. So I'm just going to write a constructor circled for constructors CTO, our tub, tub. And then you could actually just write in the cosmos clients. And then using Control dots it will offer to initialize and inject the field for you. So I'm gonna showing a keyboard shortcuts. You can write it in quickly. We wanted to private read-only cosmos client. And that gets initialized inside of our constructor. And then we have a private container, document container. And then we're going to see document container is equal to Cosmos client.stop container by the name shopping cart items with the colleagues on items. So if you hover over that, you'll see it asks for the database ID and the container ID. With those tool. With the document container, I can now access the database, the items done to all of the operations I need. I was making the point that I find this a bit more consistent to do with a few lines than writing in individual triggers are binding code per method. What would have happened is that I would have to write Cosmos DB. Then after the initializer Cosmos DB to have the connection string values. I'm going to do that quick generally what it would look like. That's it. So that's what our notation would look like as an input bindings. So when we looked at input bindings previously, we did it with TypeScript project, where we actually modify the function dot js JSON file. There is no function that JSON file that gets generated for us when we're using the C-sharp projects. So we actually put in the bindings and triggers right here in our function As on rotation. So it would say Cosmos DB, what is the database name, what is the collection name, and then the connection string, which of course would just point to the name in the app settings. And then we would call it a document client, client know based on the situation, you might end up using different datatypes here for getting all items. We would use the client and then the query code would look like this. All right, So firstly, we would have to find a URI or generates a URI for the collection. So I'll say URI or Factory dot create, sorry, create document collection URI, and then we give it those. So here you are seeing that I'm repeating these strings all over the place. You'd probably want to think about using a constant plus a constant for the database name or concern for the collection name, etc. Not withstanding that, we would just pass in the database ID and the collection. And then we would say I document query relative to the data type where a boat to query. And then client documented Client dot create document query. And then we pass in the URI as document query. And all these can be extended to the hub where clause is equal to even say like we're and then putting in some condition by which you wanted to find the documents. It could be that you want to find or do like a search that would come in handy for something like a search. Either way, this would return all of the documents and then we'd say while the query has more results, we would query through them and then potentially populates at least that we would intend to return. So we had that list of top. I've removed that also. And I'm just using it here locally. So you would actually just add each item to the list. That was what you were doing, were not taken that approach I'm just showing you because it's good to know your options. So I'm going to comment that out, and I'm going to comment this out. So at least when you look at the source code later, you can see that this would have corresponded with that code. However, because of injected the client on top, I can now go straight and C. So firstly, I have my list generated here are declared. And then I say var items is equal to document container dot get item query I iterator relative to the type shopping cart item. This is going to be as gotta go off and just get everything and they will see awhile items has more results. We would just go ahead and add. A three-inch tool or a new list. Alright, so var response is equal to items that are read next async. And then we just drop that range into our drop-down list, into our list. And then that's what we'll return. Come to think of it. I could probably even this line charts. Let me see if this would work. And C items dot read. Next is seeing dots. Okay, so it's an asynchronous soft course I would love to our width. Well then if I do that our weights, then I would be able to do a to-do list on what is returned. There we go. So I probably don't even need all of this anyway. Two lines. You see, sometimes you write something and then you realize I could have written it a bit better. So I'm going to comment that out for now and then we can come back and test and validate if my refactor works as expected. But you see goal to get the items and then we get them is in items dots read makes the async, which basically the result isn't the items who were up that's in our briefs or in parentheses so that we can call the two list on whatever is returned. All right, that is what we're returning when we tried to get all. So let's move on. The gets by ID. So here is where it would be like for like based on what we did previously, where we actually had that binding, where we looked for the ID for the partition key, which in this case is category. So here I've modified the route to say give me ID and partition key. Then those are the values I'm going to use to run my query. I could actually have used the same kind of binding that would've used previously inside of this method here. That would actually help us to get the one, the one record that matches this ID and partition key. So here's an example of what that would look like, that binding code. So Cosmos DB, we have the same database name, collection name, and connection string. And then I went to tell it that the ID is bound to the ID coming in from the parameter here. The partition key similarly is bound to the partition key coming in here, and then that is going to be scattered. Go to Cosmos DB, do the lookup and store the result instead of shopping cart item. And then once we have shopping cart item, you wouldn't be as easy as returning bulky object results, the shopping cart item. But like I said, I prefer to kind of consistency because this, this binding code is going to look almost entirely different from this binding code. And you have to be very careful and I'm not saying it's bad, I'm not by any means bashing the code. I'm seeing that it's not my preference, but I am showing you what your options are. So I'm going to comment this out so that the PCAOB it as a point of reference. But what I have done here is to say var item is equal to our weights document container, read item async. So get item query I iterator. That's all we get to our collection. This one is one you want, one will want a shopping cart item and I give it the ID, and I give it the partition key, which is category. Of course, I likewise extended the parameter listing to include the ID and the category. And I just did a route that I point to note to have the ID slash category on it. When you're trying to get, you should pass in the ID of what you're getting, as well as the category name. You can talk with glucose. That is what we do. And then the item is actually going to come back with a status code. All right, so it's going to say if the status code is equivalent to HTTP status code, phone, then we'll return on not phone results. All right, Nice and simple. Otherwise we go ahead and return the old key results are ok. Object results with item.name resource. So you can see that erupted in a try catch because I was doing something wrong when I wrote this and I needed to see the exception. So you can't do that if you need to, but it's really and truly not necessarily. So I went to remove it just so I can simplify the code and show you exactly what is required. It so that's whole. You would look up one record from the Cosmos DB data store. Now let's review what happens when we do our Create. Alright, so what we're doing here really and truly is seeing that we want to do the same kind of bindings. So of course I commented it out, but if you look at it, it's the same kind of binding, except this one would be an outward binding on output bindings. So this is an input binding. This is an output binding because then it's going to output something to the Cosmos DB. So essentially it is getting the Cosmos DB collection database, we know all of that. But then the last parameter here. Is of type I Async collector relative to the datatype. We know we want a process and then the name of it. So let me put it in the same line so it looks a bit more uniform. That is the datatype we have to use because we're operating asynchronously. If you weren't doing it asynchronously, then we'll just do an Alt variable. You cannot use old verbals instead of an asynchronous method, we have to use the Async collector relative to the datatype and then shopping cart item. And then later on in the code, after we get the shopping cart item and everything, would actually have to do something like this to add it to the collection. So let's see. Shopping cart items ALT. So a car to this name that should have been shopping cart items. I'm just naming that because we know it's an ALT. It's all putting to the Cosmos DB, right? So shopping cart items, not add async, and then we will pass in the item that it should be adding. That's really what would happen on the end of that operation. The commented old code, much as a commented out colder it. What we are doing however, is we're seeing, of course we know we get the data from the body, we parse it accordingly, and then we create a new item of type shopping cards where we pass in the name and the category that came in through our request. And then we see document quantity and r dot create item async. We give you the item and then we'd give it the partition key value. The relevance of this partition key value is that we are using the category that was sent over as the data, right? So when I said that category, when you're creating this item and that category, we're creating it to our setting that value. So when we did our blog example where I'd use the ID for both ID and partition key. In this case, I'm showing a more practical example. Now, you'd really want to slice it based on something like category or groupings. We're using the category as the grouping. Then once you do that, we just return the old key object results with the item accordingly. So either one of those lines would work relative to the uncommented hard to come into adult court. Let us look at the update and because I removed the category from the update shopping cart item, we're getting this error, but when we get there, we can look at it. No, I'm not going to spend time looking at alternatives for the put and the delete you can go to and research those. I'm just going to press ahead and show you how I have done it. So I updated the root or a need to update the route to include the category. So I did not put in the category because the way we're doing it, we always need that partition key. So I need to extend this to the category. Of course the method is put. So we're getting the request data, parsing it over. And then I'm going to see if our items, I'm going to go and Fitch the item from the database. So we're using the same code that would've used in the GIT single, right? Well, this time the partition key is coming in from our parameter category, or we'll just assume ways. So I need to make sure I put string category. And then once it finds the item or not, it's going to check while key. If it found it, then we move on to say item dot resource. The data after it has finished, doctoral data is stored in item.name resource. That is why we returned item dot resource hopped up on not just the item. I don't know if I pointed that out earlier. Item.name, resource and then we update whatever we need to update to nowhere I wanted to update in the Collected. So if you extend it to have name on whatever it will be, item.name resource thought that property. And then we are weights document container dot upsert item async, where we send back the item dot resource for too much. So after we update item dot resource, we send it back. Then that is what we're going to return anyway, as the updated data. In the delete, we once again update the root. So we're taking the category as a partition key. And then we're just going to see a weights document container delete item is seeing. We're seeing we're deleting a shopping cart item. We want to delete the ID and that particular partition key. And then we return the result. That's basically it for this operation. All right guys, so let us test it on anything we do. Anytime we do something cool, we have to test it. I'm going to just bring up Postman and I already did a test to the post-war I created oppose and I gave you the category of fruit. Here is a successful return. We can go and get. So this is getting all and I have two records in there, the sun which and I have apples. One is in the category of food, one is in a category of fruit. So earlier when I was statistic, I was actually using needs and wants as my test. And what I noticed was that it actually returned a 500, right? So I have a cart ID card, partition keys, so that really should give me a 404. It means you cannot find what I'm looking for. A 500 indicates that there is a problem on my server which is not accurate. If you look over here, you'll see that it's actually throwing an exception because the HTTP client that is making the call to cosmos DB has a method in there that says ensure our property that says ensure success code. If it sees anything other than a 200 or one of those 200 response codes, It's actually going to run an exception. So even though it is seeing that it's a 404, it's throwing an exception which is manifesting as a 500 to the client, which we don't want. So what I've done is to refactor it up a bit. So I'm showing you the code that is being called and that is throwing an exception. My refactor would have me do a try-catch that call inside of the try. Then I return the old key object results, and then 40 catch. I'm going to say catch the cosmos exception in the form of an object called E. So Cosmo, so exception is custom exception that is being thrown by that method. And then when the status code is equivalent to and we can just use the built-in status code E norm for not phoned, then we want to return a new 404. All right, so that is my literary factor. So I'm going to restart this and try that again. If I execute this again, let me just make sure that this is running. So if I execute this again and press Send, no, there's no instance or there's no indication of an exception in the log and we're getting off 404, so that's much better. I wanted this as food. Let me try a cart. Requests that should not give a four or four or a 500, and there we go. We can retrieve the record successfully. Now let us try put. Initially I had the category in that I'm willing to take the tote and I went to just leave it at collected is true that those update our ID value. So the ID and the partition keys needed for the put, let me put both and then update. So now you'll see that this is true. So initially, we know it was false. We can see that it is true. The update works. Then let us delete. And once again, we need to put in our ID and a partition key for the delete. So let me just do that, and that's a 200. Okay. So if I tried to get again, then I only get one. So all of our code is working with our Cosmos DB. Of course, if you wanted to further validate this, you can bring up the MLA it or create looking at MLA to make sure it's there and then do all of your adjustments accordingly. Play around with it, explore, see what is possible. 31. Updates To User Interface: All right guys, so we just went through quiet our code, setting up our application to use Cosmos DB, right, so let us just quickly go through some of the UI changes that have to be made. Since we updated the roots information for some of our endpoints like we changed the get to also need the partition key, stuff like that. So of course the endpoints will need to be updated accordingly. So in our client up, we can leave the route API endpoint or the base address as is. But then in our component, I am going to make some adjustments to the update and the delete. For the updates. We had said that yes, we take the ID, but we also want the categories already have access to the item. So I'm just going to say slash, add on the category value also the same for the Delete. We're just going to add on the category. And that's really all there is to updating these endpoint calls. Of course, if you are extending it to have a detailed speech where you click on see just the detail off one record that you'll need to consider them. I mean, just delete that piece of code and show you another update that are made. Initially I had this as value equals and it was born to the bool, but that should be checked. The input type checkbox checked is equal to whatever value that is. And then we have that event and nothing changes for anything else. I'm just going to quickly test this out. Alright, so that is our user interface. I'm going to test this again pleaser to assist with Cosmos DB. And the category would be random. Create shopping cart item. There we go. All right. Of course we could always update the code that it empties out the textboxes as long as something is added successfully. Know before you would've noticed that yes, he was calling bullet. The check value wasn't showing when it was true or not, so we just updated that. So if I check this one, let me just bring up the console window so you see that it is going to call the update. There it is The put what's called successfully. So if I navigate away from that page on comeback, it's going to run a fresh query. And sometimes when you interfere with the console, it pauses. But there it is, That's the fresh query and know the check value is showing. So if I delete, we can see that the function was called, but this is not redrawing, so I can probably revisit that. But if I navigate away and come back and allow it to run the git, then we see here that it actually was deleted. So we can probably update the UI code to redraw properly after I delete. The quick and easy fix for that is to modify or onclick event callback method. So instead of doing the way that we did with the event args and then we call the method name here. We're just going to have our ad sign, open parentheses, then open and close. It's more like a delegate function and it, we're delegating it to our delete clicked item. All right, so pretty much that will just solve that redrawing issue because the way that the component loads that we were doing it before, it would actually try to reload everything before it actually registered that the items list was updated. And another thing that you might see me do is use a list instead of an array that's really as normal material effect, I just tend to prefer lists. Both areas might be more efficient. That's besides the point at this, at this time, That's how you can get that delete to redraw the page properly. But that's really it for the changes required to our client up. 32. Section Review - Create REST API: All right, so where it's another milestone and we're just going to review some of the major talking points from these activities. Major point number one, we hope onto know have been creating one function with one file or one file with a function. Nobody realized we can have multiple functions in the same class file, which in the case of us creating a rest API would be a bit more efficient in terms of not spreading all of the related endpoints or behaviors across multiple files, but just having them in one file for 1 of reference. We also looked at the fact that we can define the function names separate from the root. With our defining the rules, we could actually mimic RESTful API and mentee and the standard offers. We can get different operations like GET, post, PUT and delete included. And we also in old fit or roots with parameters that needs to be passed in which would automatically be boned relative to them being present in the method header. So those are some things that we have looked at. We didn't look at any security will do that when we go to the as your side when we are publishing this up. So all of them are running at authorization level anonymous. One of the reasons you wouldn't want to make it following shown or even master is that with each request that key has to be prisons. We looked at that in future lessons. I'm just pointing out why we remained anonymous for all of our API operations. We also looked at integrating with our data. And we, we saw that you could always add it as a binding based on whole you integrates with it, the code is going to look different. So I've included the commented snippets that would kind of go hand in hand with the operation at handwrite. However, I had took a different approach because when we use bindings, each method based on the type of binding or the we were doing. The bank may require a different data types and different ways to interact with said binding. Instead of having that, you know, all those changes all over the place, I instead went with the dependency injection option where I'm using the cosmos clients. We set up a startup class that would have bootstrapped the font that we want our cosmos client. And he's getting that from the Cosmos DB connection, which is stored in my local settings.js SON. And of course we're emulating that using the storage emulator, at least for no. And having done that, I can know just injected everywhere. So originally would've gotten public static class and static methods. I had to take all the static keyword in order to support our dependency injection with this document container, we could actually rely on. I'm more in my book consistent WE of variety in the code to accomplish certain things. Know that being said, that's why I give you the options because different folds, different strokes, different situations may call for different executions, Right? So not necessarily seeing that this is the way it needs to be done, I'm just showing you the way that I prefer. Alright. So that is really all there is to or API. We also looked at the fact that with our Cosmos DB, we have to include our partition key. And in doing that, we had to make sure that our model that represents our Cosmos DB table has Jesus On Property ID and JSON property name that matches or partition key name. The final thing that we did was just the update or UI to make sure that it was reflective of the new roads and anything else. That's really it for this module, I hope that you have accomplished a lot and I want you to feel inspired to go and try to do this on your own. You can see that the Photoshop is quite light. It's an easy way to start writing code and accomplishing things with all the overhead of two-minute project files and configurations needed with all of that said and done. See you in the next module. 33. Azure Function Deployment: All right guys, welcome back. In this section we're going to look at publishing or a 0 functions up pretty much in a 0. So there are a number of nuances that we need to pay attention to. We have spoken about the hosting models before, but we're going to take a look at them again. And we're going to explore what needs sought thought process needs to be put in when we are going to be publishing our function up. So stay tuned. 34. Publish Azure Functions: Alright guys, so let us get started with our deployment. So we're going to right-click or a function up and go to Publish. And then we get this wizard. So we wanted to publish to a 0, we select that and then we need to specify a target. I may have inadvertently mentioned earlier in this course that if you're using Windows, then you may want to use C-sharp and then you can use other languages for other platforms. The reality, however, is that.net is actually cross-platform. C-sharp will work as well on Windows as it will unmarked and Linux, that's the reality. So your function up the basic level, can be deployed on the Windows version of this function up or on the Linux version of this function up. Just the same way that Python, Java and TypeScript and JavaScript on any of that language can also be supported. You also have options to publish them as containers with Azure Container Registry, which can allow you to host your Docker image. However, I'm going to proceed with Windows. After selecting windows weekend no view or search for Photoshop app's already have a function up that way I'd created from the start, but I'm not going to use that. I wanted to create a brand new one. So I'm going to click create a new function up. And then it's going to ask me for the name. Someone just see a shopping cart hyphen API, the subscription name. Of course, this one is you may be prompted to sign in with your 0 colon, so you can proceed to do that. I'm going to leave it in one of the existing resource groups. There we go. Function of r g. If you don't have any resource group, you can actually just click New and then creates a new name. Then we have our plan types. So remember that we have the consumption plan, the premium plan, and the App Service plan. So just to recap what the plan types are and the different pros and cons of each. The consumption plan is the server, the spleen is the most affordable plan. That's pros and con is that it actually has limitations with the execution time. After about five minutes or so, the whole app will actually wind dome. So it will deallocate resources. And then when you make another request, it will take time to spool up again. So of course, for testing on for quick wins, that would be a good approach. However, for a real scenario in a production setting where you need a maximum efficiency and maximum uptime, you would probably want to use premium where it does not wind down to Resources. You will have dedicated resources at all times and you can access VNets and other resources on network. That's a network level. Of course this costs a bit more. Your other option would be fused and App Service plan. So maybe you have an app service running our website already or another API. You could actually just deploy it onto that plan. Also, I wanted to go with this consumption plan. After you select that, you choose your location, of course. And for me that will be East US tool. And then you choose your stories or cones. I only have one, so that can remain. And then we go ahead and click Create. And after that is completed, where lead back to this dialog box where we can choose to run from package file. So that's ticked and it is recommended pretty much what it's saying is that it will actually deploy more like a zip file with all of the required files, the DLLs and resource files that we need. And it will mount like an image of the app instead of the actual files being there. With old trend to understand the intricacies of it, I'm just going to leave it ticked is stated that it is recommended, so I'll follow their recommendation. Then we can hit Next. And then here they want us to decide. Do you want to publish like a regular website? Before there was CICD, this is how we used to publish. It would generate a published file. Deploy all of the files to a particular place within the network or your machine, and then he can copy and paste them to the destination. Or do you want to use CICD with GitHub? So I've already added this GitHub. If you want to, you can always go ahead and do that. You can add the source control using this option that would have been in the bottom right-hand corner of Visual Studio. After you do that tiny uploaded to GitHub, then you can probably come back here. All right, Someone to use CICD with GitHub Actions and then finish. So that means each time I check in a change, it will actually trigger deployment. Nowhere lead to our final screen where it is giving us a summary of everything that has taken, please. It's actually going to add this new file which is called a YAML file. A YAML file, which is Kelly has all the deployment instructions that GitHub is going to use to determine what needs to happen. We also get a preview off the URL that will be generated for our zeros up. So everything that we need, we can access it live through that URL. That's where our EPA will be accessible. And down here we'll see some of the dependencies Application Insights needs to be configured. Let's go ahead and do that. Remember that we would want to have that for the logging. You can actually just go ahead and click Next through this wizard and allow it to finish. Once you've done that, it was surely agreeing teeth that's configured. They're also showing you that they will replace the Azure Storage sitting is for locally we're using the emulator, but once we deploy will know that it should be using the real storage, I quote. All right, so with all of that done, and if you wanted to modify the amplifier for whatever reason, he couldn't go to More Actions, edit. And it's showing you what is required. Like I said, like I said, this will actually be triggered when we do a check-in. If I open the git changes window, you'll see that it's adding all of these changes. And then I'm going to see a deployment activity, deployment column it. Then I went to just come at all and the sink save all changes and allow it to push to the repository. We can jump over to the repository in GitHub and we can just click that little dot that appears beside the chicken or to commit numbers. So if I do that entry details, we'll jump over to that CID see CICD screen. So let's give it a few to finish up. After a few minutes, it should be done. Let us jump over to our dashboard. Let us look at our resources. So I'm going to go to all resources and I should see my new function up. There we go. Now, before we go any further with the function, remember that we do need a Cosmos DB or whatever kind of database you would have used. You do need that exist in order to have that connection ready for testing. I'm going to dump over to our dashboard and I already created the Cosmos DB, but I'm going to walk you through the process regardless. So you can go to All Services and then you just go and search for Cosmos DB. Once you have the Cosmos DB you would create. Then we're using the core SQL. No, on this screen, you choose the resource groups. So I recommend that you choose the same resource group as your function up, of course, you enter unopposed NameNode. The naming is very strict so you can't use any capital letters or anything. So I would have called the shopping cart items. Of course, that's really just an account name, so you don't necessarily have to call it that, but you give it that name if it tells you it's not available, you just try and make it a little more personalized. And then you choose a location that's best for you. Once again, I recommend choosing a location that is as close as possible to where the function up was deployed. Once you do that, you can choose whether you want to provision throughput or serverless. And if you choose serverless, obviously don't have the options below. If you choose provisioned throughputs, then you can choose to appear that apply that free tier discount. Know what I have experienced in the past is that I wasn't able to proceed with the free tier discount on the second third tray. So I'm just pointing that old. If you try with it's applied and you end up having a field deployment, you'd have to delete that resource cohere, try again. And I recommend that you do not apply. Of course, if it's your first time, then you can leave it on that. But if you choose do not apply, then it will start costing you or you can just choose serverless. You also can change the buckled policy to save costs. So since this is not a production system, you don't necessarily have to go geo-redundant. You can do local redundant backup stores so it costs less money. But if its production and I recommend that you leave that on geo-redundant, those are the real to real settings that you would have to make sure are there. And what I'll do is jump over to my already provisioned Cosmos DB account. Once you're here, you want to go to the Data Explorer? No, This Data Explorer, it looks just like what we would have experienced with the emulator. So you create your new database according to how you would have written a code, of course. And pretty much whatever you did in the military, you just repeat those steps to create your database and the container as well as the partition key inside of this one. All right, so with all of that setup, we want to make sure that we get the correct connection string. So here's a primary connection string. We can copy that and then go back to our emotion up. So I'm just going to go over to the dashboard, jump over to the function up. Then inside of the configuration, we need to add that primary connection key. Remember in our code, kudos sets up that we'd need to look in the upsetting dot JSON locally, that would have been the local settings dot JSON file. That's what you're recognized as the opposite things. However, in a deployed setting, we have to set up or Application Settings. You'll just see new application setting give you the appropriate name relative to what we wrote in the code, which has Cosmos DB connection. And then you paste that primary clinician key here. I already did that. Here it is. I added that money. Well, the Cosmos DB connection and the value. Once you have all of that setup, then you can proceed to test. To test, we can click on functions, and that will give us a list of all the functions that have been deployed in this up. If I click any one of them, Let's start with the shopping cart I attempts. Here you'll see a dashboard of all the executions. You've seen how this dashboard looks already. If you go to quarter and test, note that you cannot write code for the function here because it was actually deployed from Visual Studio. So any edits have to happen in Visual Studio. It's really just mounting it and it has the bare minimum as far as the configurations are concerned. So here it's showing you that the script israeli loading from that DLL, that's the compiled version of the whole source code, right? That is what we would see there. But if we were to go back and get the function URL, then we'd be able to get that URL, which we can just paste in the browser to test. If you want to monitor, as we saw before, you can just go to monitor and go to logs. And when that is connected and open running, you can proceed to test. So I'm just going to open in a new browser tab, sorry, and paste. Here I'm getting that empty array bracket. But then if I look in the logout, see that it called the function and there was nothing to share. All right, so that means using Postman or any other API tool, I can actually simulate the same kind of cause that I'm used to. So here is where we would have created a shopping cart item instead of posting to the local host someone to replace all of that with our new URL to our Azure Websites, send the request. And there I have it. It's a 200 OK response. So that means if I refresh this request to get all know MC in the object in the database. That's it really for publishing or function up to Microsoft Azure. 35. Deploy and Test Blazor UI App: All right guys. So in this lesson we are going to be bringing our client up up to speed. So what I've done is I've jumped over to Visual Studio. I have the program CS file, and I'm going to change this path from the local host functions path to our deployed functions path. From the dashboard. I'm just going to grab this URL as that is our new base URL. And I went to replace this line of code. So I'll just duplicate the original comment that out and paste this new address with that piece, then let's test it out. So I'm just going to right-click this project and go to debug and start with old debugging. When we navigate to our fetch data pitch, you can see here I'm getting an error. So if I go and inspect and look in the console to see wire, you're seeing that it failed to finish. That's all it's really saying. What if you look up above that, you see that it says access to fetch at that address from this origin has been blocked you to a course policy. So I've mentioned course policies earlier, of course is the cross origin resource security protocol. In other words, two things have to be on the same network or a domain or a framework infrastructure for access to be granted. What I need to do is jump back over to the functions up, go down to the course policy. There we go. Then add the URL. Once that is added on Save, then I'll come back and try again and notice there's no error. No, I have data coming from as you're alright. All right, so now that I've verified that this is working, Let's jump back over to the program.cs. And what I wanted to do is configured this, that if we're in production, we use the Production link for interviews. The deputy. Alright. So that I can just set it up one time. I went to see, give me a var B's address, creates a variable called a base address and set it to be this new URI. Or rather let me set it to be this string, which is our local. And then I went to say, if builder dot host environment, it is, well, this is the development address. I'm going to say is production. This is a nice, easy way to check if you're in production when you're using a blazer WebAssembly up, then I'm just going to see the base address is equal to this string, which is our published address. Then I can have one line where I see blurred out services issue behind me. It's address is new URI off the base address value. With that done, it will be dynamic. There might be more elegant ways to do that, but, and all this will work. Let us publish our up. I'm just going to right-click. Go down to publish, publish. There we go. Then I'm going to create a new profile. So the profile that's there is already for the full motion up, but this time I'm creating a new profile. I'll choose a 0 and I'm going to publish it to another app service. And it's going to ask which UPS service instances here. Let's have a few already. I'm not going to use any of those are probably can. That's fine. I'm just going to do a new one. So I'm going to create a new one. This would be the App Service plan. The App Service plan is pretty much the b's. Then you deploy to that BCE, this App Service itself. I'm going to create a new hosting plan. I'm going to call it a shopping cart list and put it in the same resource group as our function up. The importance of assigning stuff to the same function, to the same resource group would be that if you no longer wanted the up and the resources for this up, you can just delete the resource group and everything in it. And that's a nice, easy way to clean up. I'm just going to put these in there also. And I'm going to place this on a free plan. By clicking New. I choose my location and then I can go down to the free plan. But of course, based on your context, you don't want to choose the plan that best suits your needs. I'm going to go with free for this demo. We're going to click Create. When that is done, I can just go ahead and hit Next. And I will publish this with the CMC ICD setup as the function up. Just finish. Alright, and then this summer pages letting me know what depending URL for my app is going to be. At that point, I can just go to get changes and say setup deployment for blazer, Comments All and sync. And then if you want again, jump over to GitHub and watch that deployment. It's done. So let's jump over to the published sides. Remember, we can just get that URL from the published profile or go to the board TO dashboard and get it from the newly deployed up. I just scroll to fetch data just to make sure it can communicate. It cannot. Can you guess and tell me why? Well, if you guessed that the course policies not setup, then you guessed correctly, right? Here we can see it. It's still compatible course. Why is it complaining of both cores? Because that's a brand new domain name separate from the one that we had used initially, right? I have to go back over to the function up, jump over to the course policy configuration, and this new URL. Let me just make sure I'm using the format. Save and enjoy the fruits of our labor that is publishing our clients set up, which is communicating with our function up. There are other things that we can do with this other configurations that we can make along the way to meet this more seamless. I will discuss them later on, but for now, we have successfully published both the API and our client. 36. Setup Security: All right, so now we're going to discuss security. Let us see that we hover up. Yes. We want to secure the function up and we don't want to allow every n, anybody to be able to access the data or just school anywhere they want. So as you can see so far, URLs are quite open to the function up anybody who gets a hold of the URL can figure out the pattern or if they've seen it at least once, they just need to put slash API And then that endpoint to be able to access that. No. There are different ways that you can secure an API at the very basic level, you can put on like an off key or request that an off-key be presented in the header. The current industry standard is using JWT or JSON Web Tokens. But the downside quota and coaches that you would actually have to go and configure your function up. Look for these tokens, are these keys in every single request. And you'd have to remember to do that for every single function. That can lead to overhead, especially when you have multiple functions and multiple operations that you're supporting. Another way would be to use the keys that we would have looked at, the UPC keys where we can sit it from default or from anonymous, function level versus admin master level. The downsides of this is that once you set the key, that key has to be present in the URL and that key doesn't change unless you command changes. It's either you're going to have to maintain GnG need to every time or you're going to have to expose that key. And the key would be present in the URL coming from our requesting client set up. Anyway, so somebody could still be able to get that information and be able to access your API. That would meet more sense if that was like at the network level of machine to machine. So if you create a function ups and you had like another server set up calling the function up. So that means you're not seeing any URL traffic, its GTP traffic going over in the browser or monitor Dr. MB, sniff old very easily. Then he had the keys will be a very quick win. However, because we have the client side up, I think it's better to kind of impose a security at a client-side level or make it seem like it's at a client-side level. So when a user tries to come into the app, they will have to authenticate here in order to get to the function. So an easy way to get started with that is to jump over to the authentication option in our dashboard for our function up. Then we're going to add an identity provider. Now this is going to be like mini version of Azure AD. If you've used Azure AD, then you know you can use different providers, ones their support OAuth2 Standard. If you are not so familiar, you can check out my Microsoft Azure for dotnet developers where we go through setting that up with azure AD, whether it's for the business or for word facing customers. But this is a very scaled down version two, that you don't even need a full understanding of Azure AD in order to accomplish what we're about to do. Firstly, we choose identity provider know each one of these might require some amount of setup. Of course, for Twitter, Google and Facebook, you're going to have two global there and register your app and all sorts of stuff. Someone to choose the path of these resistance. I'm just going to use Microsoft. Then it's just going to ask me, okay, what do I want up for distribution? Yes, storage or cones. The support an opponent types, you probably want to pay a little attention here to make sure that you are going to choose the one that is best for your operation. I can choose only the current tenant, meaning the tenant being the account that the app has been it deployed onto any user that has access to that document are easy in that tenant as far as as your AD is concerned, then they only those I can also say any Azure AD Directory. So if it's multiple tenants, I can choose that. I can see any Active Directory and personal Microsoft account, and I can see him personal Microsoft accounts only. I'm going to go with this one. Even though this one would work for probably about, once again, the Azure AD, those user management. So that's why the one that you choose is very important. Once again, if you want more details on that, you can get that told in the other course. We'll do it with any Azure AD and personal Microsoft accounts. That should allow me to login with even my personal known as your account if I had one up service authentication on settings, how do we restrict access? When I went to see, you can just read requiring author authentication ensures that everybody will need to authenticate. If you allow unauthenticated requests, then you'll need your own code. That's no problem. I will just require authentication. So once you are heating up here is just going to call the API or the function up. I require authentication. For authentication on authenticated requests. We can choose the one based on how you want your user experience. And this will redirect to Microsoft and we will store the Tolkien. So I'll just click add. With all of that done and added or perceived. We get the screen and null. We should be able to enjoy some amount of authentication. So if I try to go to the URL for the function up directly is just going to redirect me to authorize or authenticate. If I wasn't already logged in, then it would ask me to put in my credentials and then it would allow me access here. It's seeing that I'm requesting permission to this API. And I would want to accept that. After accepting that, it will give me consent. And then I can go to the function up no window and said that this activity is that it will stop working in our blazer up. That's something new for a number of reasons. You're going to see a corps are, and what's happening is that it's saying that it cannot redirect to the login URL because it's me redirected from the API URL, from the website URL. So there's a lot more configuration that needs to hop in there. And that's not going to happen in this particular lesson. I really just wanted to show you how you would configure and even secure the functions up. But then to get blazer ready for a 0 AD configuration is outside of the scope of the Azure Functions content of this course. But I'm just showing you some of the ramifications on things that you need to look out for. So you can always go back to your API and you can actually modify different aspects of this authentication that you've added. You actually added as your AD in a very express suite, right? So instead of going to Azure AD and city02 manually, just by adding authentication on this up, you'd have actually told it that you want Azure AD to secure the up. And you can go back, you can edit the authentication, you can just disable it. If you don't want it to or enable, you can change the rules here based on what you need. You can also go over to the identity provider and you can click and modify other things. Although things being the different aspects of the Azure AD sittings that have been applied. Again, see the endpoints here. You can get to your client up on your tenant, up, sorry, a client and application IDs. You can get to your app ID URL. If you'd need to build out on this, you can. Like I said, though, Azure AD is not in the scope of this course, so I'm not going to go into too many details of what is possible. So that's really how easy it is to secure your function up on pretty much any that you deploy. An App Service that you deploy can be secured just as easily. 37. Section Review - Publish Azure Function App: All right guys, so in this section we covered quite a few things. We looked at how we can publish our Azure up that we've been working on in Visual Studio. We also looked at some of the changes that will be needed for both the published up as well as any other client up that mean needs to interact with it. We have setup published proof us for both the function and our client app, which was really just there as a demonstrative I'm piece but it's not really the focal points, so it didn't spend too much time on it. But we were able to deploy them using CICD pipelines courtesy of GitHub and see how your mix everything easy and seamless. We also looked at some security options for our functions up using Azure AD and the whole everything can just flow. At the end of this, you should be able to get your 0 functions up, up onto Azure and ready for production. 38. Conclusion: Guys, congratulations for making it to the end of this course where you have learned all you need to know about functions. In this course, we covered setting up a 0 function using the portal. How we can write functions right here in the portal, the different types of triggers that can be used. So we can have a blob triggered function versus a timer triggered one, or maybe the most commonly used one which is an HTTP trigger a function. You also looked at the fact that you have different deployment models. You have the serverless one versus the premium plan versus the consumption plan, each one with its own benefit. We also went through and built our own suite of functions which we deployed in its own up and apps like arrest fully API. Here we were able to not have to write much code and not have to worry about any servers and configuring IIS. All we did was write a bunch of functions that interact with a very simple Cosmos DB database can be called from any client up. Once deployed on a 0, we looked at how we can access them by the Internet whole weekend setup course policies to make sure that they are fully accessible. All we need is that asterisk to say anybody can access it. We also looked at how we can secure it using Azure AD. Notwithstanding the very specific example that we went through, we also explore different options of functions like durable functions. We looked at the different kinds of bindings that can be used. And we looked at all the different tools that can come together to help us get to where we are. So we looked at emulators, we looked at the core tools, we looked at Visual Studio Code vs Visual Studio. Whatever your platform is, there is a tool that caters to you with all that said and done. Thank you for coming along with me on this journey. And I can't wait to see what you produce.