Thursday, July 23, 2020

Excluding Teams from Office Deployment

Recently I received a request from one of my readers on how to tailor Office deployments and to be specific, how to exclude Teams from deployment configuration file.

Let’s navigate to from the browser. This is the place from where administrators can manage and deploy office products and subscriptions.

You can see that there are two options on the page: 
  • Create a new configuration 
  • Import your configuration

Here I am going to create a new configuration, but if you have existing configuration, you can import that too and update it based on your business needs.

Once you have entered the Deployment Settings page, there are many options which need to be configured.

Architecture Selection

Select the architecture for which we are creating a deployment:

Office Suite Selection

Next is to select the Office Suite:

Version Selection

Next is to select the version which we want to deploy:

App Exclusion

Now comes the most important part, wherein we are going the exclude the apps which we do not want to export as part of our deployment script:

Language Selection

Next mandatory parameter is the to select the primary language:

File Format Selection

We are almost done. Final step is to export this newly created configuration, and that can be done by clicking on the Export button on top of the page.

As soon as Export button is clicked, another dialog will pop up asking for the file format:

Accept Licensing Terms

Next is to accept the license agreement and provide a name for the configuration file:

Export Configuration

Click on Export and deployment file will get downloaded to your machine. Lets open the file and have a look at the configuration settings:

In above image, you can see that Teams is excluded and will no more be part of our Office deployment. 

Happy Deployment.

Monday, July 20, 2020

Acessing Azure DevOps By Using PAT

You may have come across a requirement, wherein you needed to update Azure DevOps objects programmatically and it is obvious that there must be some authentication mechanism which has to be in place.

Now, there are many methods one can use to authenticate, but for this post, I’ve specifically chosen personal access token. PAT, which is short for Personal Access Token is a way to provide an alternate password to authenticate to Azure DevOps.

To understand more about how to generate this token and how to utilize this, let’s follow certain steps and make a successful REST API call. More...

Wednesday, July 15, 2020

Made debut as an International Speaker

Today, I've something good to share. Recently I was given an opportunity to talk in one of the technical conference named 'Lightup Virtual Conference', which was a fund raising event to support UNICEF in order to take a stand against COVID19 and it was organized by C# Corner and The Tech Platform. 

I spoke on a topic 'Azure Bot Services Utilizing LUIS Capabilities' which was a very wonderful experience. Hope recording of the same will be available soon.

Sunday, June 7, 2020

Generating Client Code from OData Metadata

Sometimes when we need to call APIs, we are just given the information about the entities in the form of OData metadata. Now when we are not aware about possible methods and their parameters which are exposed for use, it becomes very cumbersome and time consuming to call even a single method.

Till some extent, we can still manage our life for Get methods as they are easy to guess using entity names, but when it comes to PUT/POST/PATCH, things are not straight forward.

And here comes the need to generate classes with given metadata. Let’s have a look, on how we can generate these client-side proxy classes.

Install the required Visual Studio Extension

I’m using Visual Studio 2019 but the equivalent features are available in other versions too. Navigate to Manage Extensions dialog and search for OData Connected Service as shown below and install it.

Using this tool, one can easily generate classes for all the entities and complex types which are mentioned in the metadata.

Generating Proxy Classes
Next is to open the project in Visual Studio, inside which proxy classes have to be generated. Once the OData Connected Service extension is installed successfully, right click on the project and selected Add Connected Service as shown below:

Next is to select OData Connected Service as shown below:

Next is to configure the endpoints, but before that, get ready with metadata in the form of an XML file. Here is the gist of how metadata looks like:

Let’s browse the metadata file as shown below:

Click on Next and select all the required entities for which schema is to be generated as shown:

Click on Next and select all the required Function/Action which needs to be imported as shown below:

Clicking on Next will take you to next screen wherein you can mention the class file name, in which all the generated code would be saved. Here I am taking the class name as RetailReference as shown below:

Now if you wish to place all the generated code in respective separate-separate files, rather than pushing everything into the single file, then this setting can be done by clicking on the Advanced Settings link as shown in the above screenshot, which will open up below options:

There are a few more options under Advanced Settings, which can be utilized based on our coding guidelines.

Click on Finish and you will notice that all the entities are added to solution explorer as shown below:

We are all set to utilize our classes. Happy learning!

Friday, June 5, 2020

Making a call to Retail Server APIs

This article will talk about how to make a call to Retail APIs(non anonymousand what all information is required to get the response.

I started by generating the access token using username-password flow and obviously the client id as shown in below image:

Then I tried to make a call to an API using Postman as shown below:

And here is the 401 Unauthorized error ☹ and the reason is - Microsoft_Dynamics_Commerce_Runtime_DeviceTokenNotPresent

After spending hours, I got to know that Retail APIs can’t be called just by passing the access token. In order to make API call successful, there is one additional information ‘devicetoken’, which needs to be sent. 

Now where to pass this information?

Well, fortunately I was able to figure it out. This devicetoken has to be passed as an header while making API call as shown below:

Once device token is passed, I received the expected response from the API. 

Hope I saved your hours. Enjoy troubleshooting!

Friday, May 29, 2020

Providing Admin Consent to Azure Registered Application

It has been a while since I drafted anything. So, all these times, I was busy learning new things, which includes Dynamics 365, Graph API, MS Teams, some of the Azure services and much more. 

Basically, this entire tenure was full of ups and downs where some things went very smooth and some things took many hours to get sorted.

But now, I am back with so much on my plate and of course my next few posts would be majorly on troubleshooting part and how-tos. May be, there would be similar solutions which you may find on the internet, but I would still love to add it on my own blogger with my own findings for my own future references.

With all this, let us get started with our first troubleshooting tip.

The requirement was about adding a new Task in Planner, which is part of Office 365. Now for any adding Task, we have to traverse through Groups, then Bucket and then inside any defined Bucket we can create a Task

In order to perform all these, authentication and token generation part have to be in place because using JWT tokens only we are going to interact with our application, which is registered via App Registrations under the Azure portal. Here is the snapshot of how API permissions look like after registration of app: 

Now with everything in place, when I tried to launch my application from Visual Studio, I got the below error:

Which means that in order to access the registered application, few consents are required. Now how to provide this consent and what this consent is required for?

If you will closely see the first screenshot having API permissions, you will notice that for performing the Read and WriteAll operations under Microsoft Graph, Admin consent is required but it is not granted [Just to brief you on these Read/WriteAll, these are required in order to perform any operations under Planner].

So, how to provide this consent and who will provide this?

Well, this consent would be provided by the admin of the application by simply hitting an URL in his browser. This URL would contain TenantId as well as ClientId as follows:<TenantId>/adminconsent?client_id=<ClientId>

So, as soon as proper URL is entered into the browser, below dialog will pop up: 

As soon as Admin clicks on Accept button, all the consents would be granted and this can be verified by going back to API permissions page on the Azure portal as shown below: 

Happy troubleshooting.  

Thursday, April 16, 2020

Jupyter error - No module named ‘selenium’

Recently I installed Anaconda to learn more about it and the first thing I was about to try was opening a web page automatically using Selenium.

So, to perform this, I used Jupyter and tried to import Selenium webdriver. Till here, everything went well, but when I ran my code using Jupyter Notebook, I got an error: ‘No module named ‘selenium’.

The strange thing is, I got an error, although I have got Selenium installed on my machine using pip with below command: 

pip install selenium.

Now what could be the reason?

So, to analyze it further, I wrote the same Python code in Visual Studio and ran it. It worked perfectly alright.

So, I just thought to give a try to check the version of Selenium and first I tried with pip as shown below:

As the above message says, it is already installed and didn’t complain anything. So, next I thought to try with Anaconda command prompt as shown below: 

Did you notice that rectangle marked with orange? 

Yes, that was the culprit who was not allowing my code to work. There was a difference in versions and as Jupyter was launched from Anaconda, it was not able to get the correct version.

Once above code ran, I was successfully able to run my below code:

Hope this trick will save you hours.