Running a Power App as Edge Extension

This post has been republished via RSS; it originally appeared at: New blog articles in Microsoft Tech Community.

To get started you do need a sample, the demo used to get you started is a solution to capture interesting pages in your browser to read later. I am aware that there are dedicated solutions for that, but sometimes I suffer from the not invented here syndrome and want to build something myself…

 

The final result will look as follows: 

edge-plugin.gif

 

There are three building blocks that make up the logic: 

  1. Edge Extension
  2. Power App 
  3. Flow 

If you count the ToDo there is a fourth one, but we are only using a ToDo list to store data from the Flow so there is no real logic required on that side.  The goal is thus to run our Power App as an Edge Extension.

 

Edge Extension

Up until this scenario I never had build an Edge Extension so I resorted to Google to figure out the steps required. As it turns out there are quite a few tutorials and the process itself is straightforward. You have a manifest.json that describes the extension and a set of files like icons, html and JavaScript. You can walk through the Create an extension tutorial to get a feeling of all the components. 

 

Our Manifest itself is straight forward. We will be using two files frame.html and frame.js, those must be in the web_accessible_resources. We also will use a contentscript.js file that is added as a content_scripts. Since we will be authenticating users, we need to have access to the cookies from our extension. Adding the "permissions": ["cookies"] fixes that for us. The full manifest.json would look something like this:

 

 

{ "name": "Power App", "version": "0.0.0.3", "manifest_version": 2, "description": "Loading your custom Power App into the browser.", "content_scripts": [{ "js": ["contentscript.js"], "matches": ["*://*/*"], "all_frames": true }], "web_accessible_resources": [ "frame.html", "frame.js" ], "permissions": [ "cookies" ] }

 

 

 

The contentscript.js will load our frame.html as an iframe. There is some logic to make sure recursive frame insertion will not happen and will only load if the iFrame is not redirecting us to https://login.microsoftonline.com.

 

The login.microsoftonline.com page will fail to render if you are injecting an Iframe due to security concerns so make sure this is in place. Finally we add some query string parameters origin and title to our loaded frame.html. That way we can capture the page we inject our iframe on. This captured data can be used by our Power App later on. The full contentscript.js looks as follows:

 

 

 

// Avoid recursive frame insertion... var extensionOrigin = 'chrome-extension://' + chrome.runtime.id; if (!location.ancestorOrigins.contains(extensionOrigin) && !location.origin.startsWith("https://login.microsoftonline.com")) { var iframe = document.createElement('iframe'); // Must be declared at web_accessible_resources in manifest.json iframe.src=chrome.runtime.getURL('frame.html?origin=' + encodeURI(location.href) + "&title=" + encodeURI(document.title)); // Some styles for a fancy sidebar iframe.style.cssText = 'position:fixed;top:0;left:10;display:block;' + 'width:300px;height:530px;overflow=hidden;z-index:1000;'; document.body.appendChild(iframe); }

 

 

 

The frame.html file can be an almost empty HTML file or contain as much styling as you want. I opted for a clean html file with the magic happening here. 

<body> <div id="iframeplaceholder"></div> <script src="frame.js"></script> </body>

I opted for a frame.js file as inline JavaScript is not allowed in an extension. By adding the  frame.html  and  frame.js to the web_accessible_resources this no longer is an issue.  The frame.js is used to get the query string parameters, add them to our Power App play URL and set that URL as the Iframe. Make sure to update the iframe.src URL to reflect your Power App ID.

 
const params = new URLSearchParams(window.location.search) var iframe = document.createElement('iframe'); iframe.src="https://apps.powerapps.com/play/xxxxxxx?source=iframe&title=" + encodeURI(params.get('title')) + "&url=" + encodeURI(params.get('origin')) var myFrame = document.getElementById('iframeplaceholder'); myFrame.appendChild(iframe);

 

This are all the steps required to run a Power App in your Edge Browser. However, make sure to reload your extension when side loading. Changes might not always trickle down as some of the content is cached. My best experience was to reload the extension each time I changed something in the JS or HTML.

 

Power Apps

Once you have your Edge Extension working the Power App can be running whatever you like. There is a bit of a chicken and egg issue as you need the Power App ID for your Edge Extension to work, so you could start with creating an empty app and save it to get the ID you need. Or you could update the Edge Extension once the App is ready. To create an application that can create something in To Do I picked a canvas app. There are no ‘real’ requirements on the app side of things, but keep in mind the sizing of your iframe when designing parts the app itself. There is a To-Do connector you can use, but that is slightly limited. I picked the Power Automate connector to have more control over the options passed on. 

 

The Power App itself only uses the two parameters that are passed on, and in return passes those on to the Flow. Passing on parameters can be done using the Param option in the function window. My call to the Flow is linked to the OnSelect and runs as follows:

'PowerApp-ToDo'.Run(Param("title"), Param("url"));

That is all there is to our app; a single button that calls our Flow. 

 

Flow

The last component is the Flow, it uses the Power App trigger and a single action to create a new ToDo. Since we are passing the title and URL we can use those in the Flow as input parameters for our action using the following expression: @{triggerBody()['Addato-do(V3)_Title']}. The full flow looks like this:

 

flow-todo.png

 

Fusion Development

Now I understand that some of these steps could have been done writing actual code instead of using the Power Platform. However the fact that I could click together a few components and have them interact with my Edge browser really appealed to me :rocket:. Building a plugin to capture specific data from your browser has become way easier. Something that fits within the fusion development approach you hear a lot about. So instead of diving head first into complex code I decided to play around with the Power App approach and I must admit I am not disappointed. The next version should have a little more logic in place to prevent the plugin itself opening in multiple iframes on a page, but other then that I had this demo working quicker then expected. 

REMEMBER: these articles are REPUBLISHED. Your best bet to get a reply is to follow the link at the top of the post to the ORIGINAL post! BUT you're more than welcome to start discussions here:

This site uses Akismet to reduce spam. Learn how your comment data is processed.