component-managed-updates.md

Managed component updates

Manage component updates and additions with preconfigured default answers.

Overview

When working with copier-based components, you often need to provide the same answers repeatedly. This feature provides three ways to streamline this:

1. CLI --data arguments: Pass answers directly on the command line
2. Data file: Store default answers in a YAML file (follows copier's data_file convention)
3. Automatic discovery: Data files are automatically discovered in priority order

CLI usage

Component Add

Use dr component add to add a component repository to your application. This will launch an interactive wizard unless you provide answers via CLI arguments or a data file.

To add a component with specific answers:

dr component add https://github.com/datarobot-community/af-component-agent.git \
  --data base_answers_file=.datarobot/answers/base.yml \
  --data llm_answers_file=.datarobot/answers/llm-llm.yml \
  --data use_low_code_interface=false

To add a component using a specific data file:

dr component add https://github.com/datarobot-community/af-component-agent.git \
  --data-file .datarobot/.copier-answers-defaults.yaml

### Post-add environment validation

After the CLI adds a component, it validates your `.env` file.

- If validation passes, the command continues.
- If validation fails, the CLI launches `dr dotenv edit` so you can add required variables.
- If the edit flow fails or you exit early, the component is still added and the CLI prints a warning.

This behavior helps keep your application configuration in sync when a new component introduces new required environment variables.

Component Update

Update a component with specific answers:

dr component update .datarobot/answers/agent-writer_agent.yml \
  --data base_answers_file=.datarobot/answers/base.yml \
  --data llm_answers_file=.datarobot/answers/llm-llm.yml \
  --data use_low_code_interface=false

Update using a specific data file:

dr component update .datarobot/answers/agent-writer_agent.yml \
  --data-file my-custom-defaults.yaml

This is equivalent to the copier command:

copier update -a .datarobot/answers/agent-writer_agent.yml \
  --data base_answers_file=.datarobot/answers/base.yml \
  --data llm_answers_file=.datarobot/answers/llm-llm.yml \
  --data use_low_code_interface=false

Extra flags

By default, python warnings emitted by copier are suppressed. You can use --verbose or --debug flags to show them.

Data file

Naming and location

Following copier's data_file convention, the default filename is .copier-answers-defaults.yaml.

Discovery priority order:

1. Explicit path via --data-file flag (highest priority)
2. Repository root: .datarobot/.copier-answers-defaults.yaml
3. User config directory: ~/.config/datarobot/.copier-answers-defaults.yaml
4. Legacy location: ~/.config/datarobot/component-defaults.yaml (backward compatibility)

Format

defaults:
  https://github.com/datarobot-community/af-component-agent.git:
    base_answers_file: .datarobot/answers/base.yml
    llm_answers_file: .datarobot/answers/llm-llm.yml
    use_low_code_interface: false

How It Works

1. When you run dr component add or dr component update, the CLI:

   • Reads the copier answers file to determine the component's repository URL
   • Looks up defaults for that repository in component-defaults.yaml
   • Applies those defaults automatically

2. CLI --data arguments always take precedence over configured defaults

3. Any questions not covered by defaults or CLI arguments will be prompted interactively (unless you use copier's -A flag, which is included in dr component update)

Recommended Setup

For team-wide defaults (committed to the repository):

Create .datarobot/.copier-answers-defaults.yaml in your repository root:

defaults:
  https://github.com/datarobot-community/af-component-agent.git:
    base_answers_file: .datarobot/answers/base.yml
    llm_answers_file: .datarobot/answers/llm-llm.yml
    use_low_code_interface: false

For personal defaults (not committed):

Create ~/.config/datarobot/.copier-answers-defaults.yaml:

defaults:
  https://github.com/datarobot-community/af-component-agent.git:
    agent_template_framework: langgraph

Example Workflow

With the repository data file shown above, running:

dr component add https://github.com/datarobot-community/af-component-agent.git

Is equivalent to:

dr component add https://github.com/datarobot-community/af-component-agent.git \
  --data base_answers_file=.datarobot/answers/base.yml \
  --data llm_answers_file=.datarobot/answers/llm-llm.yml \
  --data use_low_code_interface=false

If you want to override a default:

dr component add https://github.com/datarobot-community/af-component-agent.git \
  --data use_low_code_interface=true

This will use the configured defaults for base_answers_file and llm_answers_file, but override use_low_code_interface to true.

You can also use a different data file temporarily:

dr component add https://github.com/datarobot-community/af-component-agent.git \
  --data-file /path/to/custom-defaults.yaml

Benefits

1. Team consistency—repository-level defaults keep answers consistent across team members.
2. Speed—skip repetitive prompts for common answers.
3. Flexibility—override defaults per command.
4. Personal customization—use local defaults without committing them.
5. Compatibility—follows copier's data_file convention.
6. Version control friendly—commit repository defaults and keep personal defaults local.

Data types

The data file and --data arguments support all copier question types:

• Strings: key: value or --data key=value
• Booleans: key: true or --data use_feature=false
• Integers: key: 42 or --data count=42
• Floats: key: 3.14 or --data height=1.83
• Lists (for multiselect): key: [item1, item2] or --data items=[1, 2, 3]
• Objects/Maps: key: {nested: value} or --data config={key: value}
• Null: key: null or --data optional=null
• YAML/JSON complex types: Any valid YAML structure in the data file

The data file uses standard YAML syntax. CLI --data arguments are parsed according to the question's type defined in the copier template.