Deploy Your Semantic Models

The buster deploy command publishes your semantic models to the Buster platform, making them available for natural language querying. This is the final step in the Buster model development workflow.

Basic Usage

buster deploy 

Command Options

OptionDescription
--path PATHOptional path to a specific model file or directory to deploy. If not provided, uses the model paths from buster.yml.
--dry-runValidate models without actually deploying them (default: false).
--recursiveRecursively search for model files in subdirectories (default: true).

Authentication Requirement

Before using the deploy command, you must be authenticated with the Buster API:

buster auth

The deploy command will automatically check for valid authentication and prompt you if not authenticated.

Deployment Process

When you run buster deploy, the following steps occur:

  1. Authentication Check: Verifies you have valid Buster API credentials
  2. File Discovery: Finds all semantic model files in the specified paths
  3. Validation: Parses and validates each model (similar to buster parse)
  4. Conflict Resolution: Handles any conflicts with existing models
  5. Publishing: Uploads the models to the Buster platform
  6. Activation: Makes models available for natural language querying

Example Output

The command provides detailed progress information during deployment:

🚀 Deploying models to Buster...

Found 5 semantic model files to process
⏳ Validating models...
✅ All models valid
⏳ Uploading models to Buster API...
✅ Successfully deployed 5 models

Deployment summary:
- New models: 2
- Updated models: 3
- Failed models: 0

✨ Deployment complete! Your models are now available for querying.

Examples

Deploy All Models

Deploy all models defined in your buster.yml:

buster deploy

Deploy Specific Directory

Deploy only models in a specific directory:

buster deploy --path models/marketing

Validate Without Deploying

Check if models would deploy successfully without actually deploying them:

buster deploy --dry-run

Troubleshooting

Common issues and their solutions:

  • Authentication errors: Run buster auth to refresh your credentials
  • Validation failures: Run buster parse first to identify and fix issues
  • API connectivity issues: Check your network connection and firewall settings
  • Incompatible models: Ensure your models follow the latest Buster semantic model schema

Best Practices

  1. Run Parse First: Use buster parse before deploying to catch issues early
  2. Use Version Control: Commit your models to version control before deployment
  3. Test Locally: Use the --dry-run flag to validate without deploying
  4. Deploy Regularly: Keep your deployed models in sync with your development
  5. Check Deployment Status: Verify in the Buster UI that models deployed correctly