This project is being archived. As part of the archival process, we're closing all open issues and pull requests.
You can continue to use this sample "as-is", but it won't be maintained moving forward. We apologize for any inconvenience.
- Introduction
- Prerequisites
- Get a key for the Text Analytics API
- Publish the LUIS model
- Register the web application
- Build the sample
- Run the sample
- Questions and comments
- Contributing
- Additional resources
This sample shows how to build a Microsoft Bot Framework bot that connects to Microsoft Graph, searches through email messages, and performs sentiment analysis on those messages by using the Microsoft Text Analytics cognitive service. It leverages the Language Understanding Intelligent Service to help your bot understand the queries you send to it.
To use the Microsoft Graph Connect Sample for Node.js, you need the following:
- Node.js
- TypeScript
- The Bot Framework Emulator
- Either a Microsoft account or a work or school account
- A Microsoft Azure Account
- A Language Understanding Intelligent Service Account
In addition to these prerequisites, you'll need to get a subscription key for the Text Analytics API), publish a LUIS model, and register a web client application. The following three sections describe how to get these additional resources.
-
Sign in to the Azure Portal.
-
Choose More services at the bottom of the left column of the page.
-
In the Filter box at the top of the More Services column, type Cognitive Services.
-
Choose the Cognitive Services option when it appears.
-
Choose the Add button at the top of the page.
-
Choose the Text Analytics API (preview) option in the right pane.
-
Choose the Create button at the bottom of the right pane.
-
Type a name for your subscription and select a plan. You may select the free tier for 5,000 transactions per month.
-
Read and confirm that you've understood the notice near the bottom of the pane, and then select the Create button at the very bottom.
-
After you've created your subscription, navigate to your subscription and select Keys under the Resource Management section of the center pane.
-
Copy the value of the first key ("Key 1"). You'll use this value in the Build the sample section.
-
Navigate to the Language Understanding Intelligent Service (LUIS) home page.
-
Create a LUIS account if you don't already have one. You'll automatically navigate to your account dashboard.
-
Choose My apps in the upper left corner of the dashboard page.
-
Choose Import App
-
In the popup, browse to the SearchBotLUIS.json file in the root directory of this repo.
-
Choose the Import button. You'll automatically navigate to the dashboard page of your new app.
-
Choose Publish App in the left menu.
-
Select the default BootstrapKey option as the value for Endpoint key, and then select the Publish button.
-
Copy the Endpoint url value. You'll use this value in the Build the sample section.
-
Select Train and Test in the left menu to go to a page where you can run test queries against your published model. Type Search Office 365 for Graph in the test utterance pane to verify that your new application is finding an intent for this query.
-
Sign into the Azure application registration portal using either your personal or work or school account.
-
Choose New registration.
-
Enter a name for the app. For Supported account types, select Accounts in any organizational directory and personal Microsoft accounts (e.g. Skype, Xbox, Outlook.com).
-
Enter https://localhost:3980/botauth/aadv2/callback as the Redirect URI (Web).
-
Choose Register.
The app overview page displays, listing the properties of your app.
-
Copy the Application (client) ID from the overview page. This is the unique identifier for your app.
-
Go to the Certificates & secrets page under Manage. Choose New client secret, give it a description and expiration date, and then select Add.
-
Copy and save the new client secret's value. This is the only time you will be able to see the client secret.
You'll use the application ID and client secret to configure the sample app in the next section.
-
Download or clone this repo.
-
Using your favorite IDE, open the .env file in the root directory of the repo.
-
Replace the ENTER_YOUR_COG_SERVICE_SUBSCRIPTION_KEY placeholder value with the Text Analytics API key that you copied in the Get a key for the Text Analytics API section.
-
Replace the ENTER_YOUR_LUIS_MODEL_URL placeholder value with the endpoint url that you copied in the Publish the LUIS model section.
-
Replace the ENTER_YOUR_CLIENT_ID and ENTER_YOUR_CLIENT_SECRET placeholder values with the application Id and application secret that you copied in the Register the web application section.
-
In a command prompt, run the following command in the root directory. This installs the project dependencies.
npm install
- Run the following command to start the development server.
npm start
-
Launch the Bot Framework Emulator.
-
Enter the messaging endpoint for this bot project: https://localhost:3980/api/messages. Leave the Microsoft App ID and Microsoft App Password values blank. These are for published bots.
-
Choose the Connect button.
-
Type a query into the Bot Framework Emulator in this form: Search Office 365 for Graph.
-
You'll see a Connect button in the emulator. Choose this button to launch a web page that displays a string in the browser. You'll be prompted to type that string into your emulator chat. Note that your browser might warn you that your PC doesn't trust the site's security certificate. Choose the option to continue to the web page.
-
After you type the string into your chat, you'll get a message indicating the sentiment score for your query.
-
Type additional queries to see the cognitive service's sentimental analysis at work.
We'd love to get your feedback about the Microsoft Graph Connect Sample for Node.js. You can send your questions and suggestions in the Issues section of this repository.
Questions about Microsoft Graph development in general should be posted to Stack Overflow. Make sure that your questions or comments are tagged with [microsoftgraph].
If you'd like to contribute to this sample, see CONTRIBUTING.MD.
This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.
Copyright (c) 2017 Microsoft. All rights reserved.