Microsoft 365 Community Column and view formatting tips, 15 Apr 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Live London Underground Line Status Bot in Microsoft Dataverse for Teams

Description

In this 16-minute Developer-focused demo, Leon Armston delivers the next stage in the evolution of his London Underground - Tube Status solution. An interactive Tube bot in Power Virtual Agents, using Power Automate, and configuring your Azure environment for Azure API Management and Azure Functions. Dataverse for Teams allows users to leverage Azure API Management service using a custom connector in Power Platform. See solution architecture, execution triggers, rules, calls and more. This PnP Community demo is taken from the General Microsoft 365 Development Special Interest Group (SIG) - Bi-weekly sync call recorded on April 15, 2021.

Demo Presenter: Leon Armston (Intelogy) | @LeonArmston

Supporting materials:
• Blog – Create a London Underground Line Status Bot using Power Virtual Agents & Azure Management API in Dataverse for Teams | https://www.leonarmston.com/2021/02/create-a-london-underground-line-status-bot-power-virtual-agents-azure-azure-management-api-in-dataverse-for-teams/
• Blog - Azure API Management connector on the Power Platform | https://powerapps.microsoft.com/en-us/blog/azure-api-management-connector-on-the-power-platform/

Learn more:
• Learn more about the Microsoft 365 Patterns and Practices community at: http://aka.ms/m365pnp
• Visit the Microsoft 365 Patterns and Practices community YouTube channel: https://aka.ms/m365pnp-videos

A

Welcome to my presentation today on creating a live london, underground line status, bot in dataverse for teams using power, virtual agents, power automate as your management api and as your functions apologies for the uh long title I just yeah. I was struggling to get the technologies in and a little about me. My name is leon anston and I'm a microsoft, 365 solutions consultant at intellij in london. In the united kingdom. You can find details below on how you can connect with me.

A

You can check out my blog leonardson.com or find me on twitter, linkedin or github. I will now outline the functionality that I want to showcase on this demo. So, in my opinion, one of the most exciting new features of dataverse for teams is the ability to leverage microsoft, azure, api management.

A

This allows citizen developers to use as your management api back custom connectors in power, apps or power automate in a data versus teams environment. So this means you can connect to using the azure api management api. You can connect with your resources ie as your functions to run powershell or code, but the good thing is that all of this is available using your current seated team's licensing.

A

So this is a great opportunity to use this exciting new technology with your customers or organization today. So I wanted to use uh power virtual agents in data versus teams to build a live interactive bot as its database for teams. The bot can only be used uh from the team's application or teams web clients. So there are some restrictions, but with your bot, you can interact with any live api, I.e, transportation, sports results, weather recipes or even those chuck norris api for quotations.

A

So you may recall a previous demo. I did on the london underground, where I created a live, interactive, london, underground line status, board that used power, automate and sharepoint view formatting.

A

So over the christmas period, much to the annoyance of my wife, I decided to spend time extending my previous london underground theme and created a proof of concept. London, underground line status box.

A

I will now use this diagram to explain the solution so watch my mouse. It consists of a power virtual agent's bot, running in a data versus teams environment for a specific microsoft team. We could see the microsoft the data versus teams environment perimeter highlighted in purple, so the user sends a message.

A

uh The trigger phrase of um yeah. Are there any delays. Power virtual agent then recognizes the trigger word, selects the correct topic, which then triggers a flow in power automate, which then connects to a customer connector and that's where it gets exciting.

A

The custom connector is then goes out to my as your tenant and connects to the azure management api instance, which then connects to an azure function.

A

The azure function then runs some powershell which, via a web request, connects to the london underground line status api to provide the line status and then, finally, that's all returned back along the tree, all the way back to the power virtual agent and it returns the line line status I'll go into this in more more detail, but that's at a high level.

A

I will now demo my solution to you.

A

So the london underground provides a uh yeah, an api and also a web page with the line status, so we're using this information uh that is then published to an api. So applications can call into this api and get the latest line status.

A

So I'll now show you my bots, which yeah you could build so I'll type, the trigger phrase of line status, but there's many other trigger phrases that it responds to. So it's glad to report. Then it asks me which line I'm interested in so I'll type, central, so that's the central line and then it connects to the tfl api via power automate via azure and yeah. It says that there's no delays on the central line, I'll type, the trigger phrase again, but this time I'll type, some random characters and obviously there's not a line.

A

That's uh called after those characters. So I then have some smarts where it's uh yeah then gives me a list of all the uh line names so I'll, select the waterloo and city line, and it will return that there's yeah no service until uh further notice. So now let me show you how this is put together.

A

So in your team's client uh on the database teams, you can access power, virtual agents, so I'm in there now and I go to chat bots and I can, when it loads. I can see my um power virtual agents bots that I have previously made in data versus team, so I select the london underground um dataverse for teams, environment and I can see my bot here called tubing.

A

I select tubing and then this will load my chat box, I'm interested in the topics. So these are the rules uh of uh yeah. The responses uh to the questions asked to the bot so I'll go to line status, which is which I configured and we can see all the trigger trigger phrases for this bot. So when these trigger phases are spoken, uh this topic is initiated so I'll go to the offering canvas now- and this is yeah the topic in logical order with all the rules.

A

So we can see the trigger phrases that we just um discussed and then we can see that there's a message and it asks a question we'll test the bot here by typing the trigger phrase again, and we can see that uh yeah power. Virtual agents um indicates in green uh where you are in the action tree. So we can see that we're up to question, which line are you interested in? And that corresponds to down here where it says? Which line? Are you interested in and we can type the central line.

A

And then we'll go further down the tree and to the response, but I just want to go back and just show you that it goes to power automate and that's where it connects to a custom connector and goes and gets a live api results. We've also got the smarts here, for if the line name is not found, then we have the option to select from a button all of the 15 london underground lines.

A

So I'm going to go to the power of the flow in power automate and show you how that's done and how it connects with a custom connector.

A

So if I edit the flow, I'm not going to go into great detail, because this is quite complex, this power ultimate flow, but the first trigger initiation is uh it's asking for the line name. So the topic provides the line, name that we typed in then. If we go to here. This is a custom connector that I uh built.

A

So this connects to azure management, api and queries um the london underground line status api. So, first of all, it's secured with a subscription key that only I know so it can't be remotely uh triggered anonymously and then it requests the line name from the first action.

A

Then that queries the api and if there's a matching line name, uh we formulate the response. uh If there's any reported delays or if there's delays expected, but if we formulate that all in markdown language, so we make thing bold, we make things uh headers, so it's nicely formatted for the users and then we return all that back to the user in the topic.

A

So I'm now going to go to azure, to yeah show all the components for this. So you will need them as your subscription to do this, but the custom connector. It connects directly to the um as your management api service that I've got. So that's a web layer that you can deploy in your azure environment and then publish your own apis to it. So you could publish an api direct from a logic, app, an app service or a function app in my case, I'm going to well.

A

I've already deployed a function app, so that's called line status. So this is a web api and we can go into the front end and see that it's a post request and that it wants to query parameters a subscription key and lightning.

A

So you may remember in the custom connector in power automate, we had two text boxes, asking first for the subscription key and secondly, for the line name, so that's yeah where they come from and then, if we go to the back end, so we can see the back end of the api over this right hand. Side is the is it's linked to an azure function app and that function app is called london tubes.

A

So if we go to london tubes in the function, app within functions, click on our function.

A

um With this will now load up the power shell that um yeah the powers uh this script so I'll just go through this we have. The first part um is uh yeah, where it's initiated by the line name. So if a line name is supplied to it that then populates this line name variable over here and yeah. That's coming from the request at the beginning.

A

Now that then feeds into um the line name, uh if statement. So that's this box here.

A

And we have the special part here is we use invoke web request to target the tfl transport for london tube status api, so, instead of power automate, you would use a uh yeah http connector, which would be a premium connector we're using powershell in both web requests and just sending over the web request there.

A

We convert that to a object in powershell.

A

So what happens is all of the line? Statuses are returned, so we need to filter it by the line name. So that's what's happening here and then. Finally, we convert it back to json format, ready for power automate to to use the next part of the. uh If statement we don't really use, but that returns all lines without filtering.

A

So if you wanted to do you have something in your power automate with all lines, that's what you would use and then finally, the last part is where we return the response back to the function and back to the custom connector and back to power automate.

A

So I'll now just show you the how easy it is to, after you've created your api to publish it to power, apps or auto power. Automate as a custom connector, so we can click um on powerapps or power automate, and then it will publish it to your instance.

A

So let me just return back to my demo and do a quick recap. So to recap, this solution uses power, virtual agents in microsoft, teams in database of teams to provide a london underground line status. Spot a user asks a question to the bot about a particular line, and then the line status topic is triggered, which then triggers a flow in power automate.

A

A custom connector is then used in the flow which connects to an azure management. Api instance that triggers an is your function. The azure function, then runs the powershell invoked web request command to connect to the london underground line status api.

A

Finally, this information is returned to the user by bar, and that's all so. Thank you for your time today and for watching my presentation below is a link to my blog and a direct link to the blog post, where you can find a full guide for implementing this solution. There is also a qr code, so you can scan this to take. You direct feel free to connect with me using the details below and let me know if you have any comments or questions regarding the solution.

A

I hope for you like me. It sparks ideas regarding the art of the possible with all that technologies. That's all from me. Thank you. Thanks.

B

Leon really really great stuff and, and it's a absolutely brilliant set of technology being demoed on this setup and and it's great to see actually how your tube line uh status line thing is evolving. It's been a dynamic, spf x report. It's now a bot, it's it's evolving, all the time across the different technologies in microsoft, restful, so really really cool stuff um and april is reaching out, certainly for getting you to demo this in a power platform uh community course as well, so this is really really really great stuff. Thank you for that excellent.

B

A

B

A