Spark Integration

Setup

Prerequisites

• Follow the instructions on the main README file in the root directory of this repository.
• Create a Spark account.
• Replace the value of projectId in the server.js file with your Dialogflow agent’s Project ID.

Creating a Spark Bot

• Log in to Spark and go to the New Bot page.
• Fill in the required fields and click the "Add Bot" button.

• On the next screen, take the value for your bot’s access Token and replace the value for sparkAccessToken in the server.js file.

Deploying the Integration Using Cloud Run

In your local terminal, change the active directory to the repository’s root directory.

Run the following command to save the state of your repository into GCP Container Registry. Replace PROJECT-ID with your agent’s GCP Project ID and PLATFORM with the platform subdirectory name.

gcloud builds submit --tag gcr.io/PROJECT-ID/dialogflow-PLATFORM

Deploy your integration to live using the following command. Replace PROJECT-ID with your agent’s GCP project Id, PLATFORM with the platform subdirectory name, and YOUR_KEY_FILE with the name (not path) of your Service Account JSON key file.

gcloud beta run deploy --image gcr.io/PROJECT-ID/dialogflow-PLATFORM --update-env-vars GOOGLE_APPLICATION_CREDENTIALS=YOUR_KEY_FILE --memory 1Gi

• When prompted for a target platform, select a platform by entering the corresponding number (for example, 1 for Cloud Run (fully managed)).
• When prompted for a region, select a region (for example, us-central1).
• When prompted for a service name hit enter to accept the default,
• When prompted to allow unauthenticated invocations press y,
• Copy the URL given to you, and use it according to the README file in the given integration's folder.

Take the value for the server URL printed in the console after the completion of the execution of the above command and replace the value for targetUrl in the server.js file.

Redeploy the integration with the updated change by rerunning the above two commands.

More information can be found in Cloud Run documentation.

You can view a list of your active integration deployments under Cloud Run in the GCP Console.