How to Integrate Google Cloud AI with CircleCI

Set Up Google Cloud AI Environment

• Create a Google Cloud account if you don't have one, and enable billing to access AI services.

 

• Navigate to the Google Cloud Console and enable the APIs you require, such as the Vision AI or Natural Language API.

 

• Generate service account credentials. Go to the IAM & Admin section, create a service account, and download the JSON key file.

Prepare Your Codebase

• Ensure your application can access the Google Cloud services. Store the downloaded JSON key in a secure, accessible location.

 

• Set environment variables in your application for accessing Google Cloud services. For example, if you're using a Node.js app, you might add environmental variables in your `.env` file as follows:

  GOOGLE_APPLICATION_CREDENTIALS=/path/to/your/credentials.json
  

 

• Add the Google Cloud client libraries to your application. For example, in a Node.js application, you might install the Vision library:

  npm install @google-cloud/vision
  

Configure CircleCI for Google Cloud

• Log into CircleCI and set up your project. If your project is not listed, create a configuration for it.

 

• Add environment variables to CircleCI for Google Cloud credentials. Navigate to your project settings in CircleCI, and under Environment Variables, add a variable with the key `GOOGLE_APPLICATION_CREDENTIALS` and the value that corresponds to the credentials JSON file path.

 

• Securely store the Google Cloud credentials in CircleCI. You can use base64 encoding to avoid storing the raw JSON, for example:

  base64 /path/to/your/credentials.json
  

  Paste the encoded string into an environment variable and decode it in your configuration like so:

  version: 2.1
  
  jobs:
    build:
      docker:
        - image: circleci/node:14
      steps:
        - checkout
        - run: echo $GCLOUD_JSON | base64 --decode > ${GOOGLE_APPLICATION_CREDENTIALS}
  

Create a CircleCI Config File

• Create a `.circleci/config.yml` file in your project root if it doesn’t exist.

 

• Define the job that builds, tests, and deploys your application while integrating Google Cloud AI. Here's a basic example:

  version: 2.1
  
  jobs:
    build:
      docker:
        - image: circleci/node:14
      steps:
        - checkout
        - run: echo $GCLOUD_JSON | base64 --decode > ${GOOGLE_APPLICATION_CREDENTIALS}
        - run: npm install
        - run: npm test
        - run: npm run deploy
  

Test and Verify Integration

• Commit the changes to your repository. CircleCI should automatically trigger a build using your configuration file.

 

• Check the CircleCI dashboard to monitor the process and ensure all steps, including Google Cloud AI integration tasks, run successfully.

 

• Address any errors by reviewing logs available on the CircleCI dashboard.