Tutorial: Create (typescript)

Create your skill project

  1. Execute the Skill generator with this command.
yo bot-virtualassistant:skill

At this point you have two different options to proceed:

Generate the skill using prompts

  • The generator will start prompting for some information that is needed for generating the sample:
    • What's the name of your skill? (sample-skill)

      The name of your skill (used also as your project’s name and for the root folder’s name).

    • What's the description of your skill? ()

      The description of your skill.

    • Which languages will your skill use? (by default takes all the languages
      • Chinese (zh-cn)
      • Deutsch (de-de)
      • English (en-us)
      • French (fr-fr)
      • Italian (it-it)
      • Spanish (es-es)
    • Do you want to change the new skill's location?

      A confirmation to change the destination for the generation.

    • Where do you want to generate the skill? (by default takes the path where you are running the generator)

      The destination path for the generation.

    • Looking good. Shall I go ahead and create your new skill?

      Final confirmation for creating the desired skill.

Generate the sample using CLI parameters

Option Description
-n, –skillName The name you want to give to your skill (by default takes sample-skill)
-d, –skillDesc A brief bit of text used to describe what your skill does (by default is empty)
-l, –skillLang The languages you want to use with your skill. Possible values are de, en, es, fr, it, zh (by default takes all the languages)
-p, –skillGenerationPath The path where the skill will be generated (by default takes the path where you are running the generator)
–noPrompt Do not prompt for any information or confirmation

NOTE: If you don’t use the –noPrompt option, the process will keep prompting, but using the input values by default.

Example

> yo bot-virtualassistant:skill -n "My skill" -d "A description for my new skill" -l "en-us,es-es" -p "\aPath" --noPrompt

After this, you can check the summary in your screen:

- Name: <aName>
- Description: <aDescription>
- Selected languages: <languages>
- Path: <aPath>

WARNING: The process will fail if it finds another folder with the same name of the new skill.

NOTE: Remind to have an unique skill’s name for deployment steps.

What files were created?

| - deployment                              // Files for deployment and provisioning
    | - resources                           // Resources for deployment and provisioning.
        | - LU                              // Files for deploying LUIS language models
            | - general.lu                  // General language model (e.g. Cancel, Help, Escalate, etc.)
            | - sample-skill.lu             // Sample language model for your skill
        | - template.json                   // ARM Deployment template
        | - parameters.template.json        // ARM Deployment parameters file
    | - scripts                             // PowerShell scripts for deployment and provisioning
        | - deploy.ps1                      // Deploys and provisions Azure resources and cognitive models
        | - deploy_cognitive_models.ps1     // Deploys and provisions cognitive models only
        | - update_cognitive_models.ps1     // Updates existing cognitive models
        | - luis_functions.ps1              // Functions used for deploying and updating LUIS models
        | - qna_functions.ps1               // Functions used for deploying and updating QnA Maker knowledgebases
        | - publish.ps1                     // Script to publish your Bot to Azure.
| - pipeline                                // Files for setting up an deployment pipeline in Azure DevOps
    | - sample-skill.yml                    // Sample build pipeline template for Azure DevOps
| - src                                     // Folder which contains all the Skill code before compilation
    | - adapters                            // BotAdapter implementations for configuring Middleware
        | - defaultAdapter.ts               // Configures basic middleware for local mode
    | - bots                                // ActivityHandler implementations for initializing dialog stack
        | - defaultActivityHandler.ts       // Initializes the dialog stack with a primary dialog (e.g. mainDialog)
    | - dialogs                             // Bot Framework Dialogs
        | - mainDialog.ts                   // Dialog for routing incoming messages
        | - sampleDialog.ts                 // Sample dialog which prompts user for name
        | - sampleAction.ts                 // Sample action which prompts user for name
        | - skillDialogBase.ts              // Dialog base class for shared steps and config
    | - manifest                            // Source of manifests
        | - manifest-1.0.json               // Manifest version 1.0
        | - manifest-1.1.json               // Manifest version 1.1
    | - models                              // Data models
        | - skillState.ts                   // Model for storing skill state
        | - stateProperties.ts              // Constants for state property keys
    | - responses                           // Classes and files for representing bot responses
        | - MainResponses.lg                // LG templates for MainDialog 
        | - SampleResponses.lg              // LG templates for SampleDialog 
    | - services                            // Configuration for connected services and service clients
        | - botServices.ts                  // Class representation of service clients and recognizers
        | - botSettings.ts                  // Class representation of configuration files
    | - appsettings.json                    // Configuration for application and Azure services
    | - cognitivemodels.json                // Configuration for language models, knowledgebases, and dispatch model
    | - index.ts                            // Initializes dependencies

You now have your own Skill! Follow the Deployment steps below before you try and run the project as deployment creates key dependencies required for operation.