Skip to content

start.md

Latest commit

 

History

History
350 lines (233 loc) · 11.2 KB

start.md

File metadata and controls

350 lines (233 loc) · 11.2 KB

dr start - Application quickstart

Run the application quickstart process for the current template.

Quick start

For most users, getting started is a single command:

# Run the quickstart process (interactive)
dr start

The command automatically detects your template's configuration and either runs a custom quickstart script or launches the interactive setup wizard.

Note

First time? If you're new to the CLI, start with the Quick start for step-by-step setup instructions.

Synopsis

dr start [flags]

Description

The start command (also available as quickstart) provides an automated way to initialize and launch your DataRobot application. It performs several checks and either executes a template-specific quickstart script or seamlessly launches the interactive template setup wizard.

The command streamlines the process of getting your DataRobot application up and running. It runs the following steps in order:

  1. Starting application quickstart process—displays the quickstart flow.
  2. Checking DataRobot CLI version—verifies that your CLI version meets the template's minimum requirements (from .datarobot/cli/versions.yaml). If not, prompts to run dr self update.
  3. Checking template prerequisites—verifies that required tools (e.g. Task, Git) are installed.
  4. Checking repository setup—verifies you're in a DataRobot repository (contains .datarobot/). If not, the command launches the interactive dr templates setup wizard; after setup completes, dr start runs again in the cloned directory.
  5. Finding and executing start command—searches for a start command in this order:
    • task start—if the Taskfile defines a start task, it runs immediately (no confirmation).
    • Executable quickstart script—if an executable script in .datarobot/cli/bin/ matches quickstart*, you're prompted to run it (unless --yes is used), then it executes.
    • Neither found—if you're already in a repository but no start command or script exists, the command shows a message and exits. It does not launch the setup wizard in this case.

This command is designed to work intelligently with your template's structure. If you're not in a DataRobot application template repository, it launches template setup; once in a repository, it runs task start or a quickstart script when available.

Aliases

  • dr start
  • dr quickstart

Options

  -y, --yes     Skip confirmation prompts and execute immediately
  -h, --help    Show help information

Global options

All global flags are also available.

Start command detection

The dr start command looks for a start command in the following order:

  1. task start command (highest priority)—If your template's Taskfile defines a start task, it will be executed automatically.
  2. Executable quickstart scripts—If no task start is found, the command searches for executable quickstart scripts.

Quickstart scripts

Location

Executable quickstart scripts must be placed in:

.datarobot/cli/bin/

Naming convention

Executable quickstart scripts must start with quickstart (case-sensitive):

  • ✅ quickstart
  • ✅ quickstart.sh
  • ✅ quickstart.py
  • ✅ quickstart-dev
  • ❌ Quickstart.sh (wrong case)
  • ❌ start.sh (wrong name)

If there are multiple scripts matching the pattern, the first one found in lexicographical order will be executed.

Platform-specific requirements

Unix/Linux/macOS:

  • Script must have executable permissions (chmod +x)
  • Can be any executable file (shell script, Python, compiled binary, etc.)

Windows:

  • Must have executable extension: .exe, .bat, .cmd, or .ps1

Examples

Basic usage

Run the quickstart process interactively:

dr start

If a quickstart script is found:

DataRobot AI Application Quickstart

  ✓ Starting application quickstart process...
  ✓ Checking DataRobot CLI version...
  ✓ Checking template prerequisites...
  ✓ Checking repository setup...
  → Finding and executing start command...

Found quickstart script at: .datarobot/cli/bin/quickstart.sh

Press 'y' or ENTER to confirm, 'n' to cancel

If you're not in a DataRobot repository, the repository step will trigger template setup:

  ✓ Starting application quickstart process...
  ✓ Checking DataRobot CLI version...
  ✓ Checking template prerequisites...
  → Checking repository setup...

Not in a DataRobot repository. Launching template setup...

The command then launches the interactive dr templates setup wizard. After you select and clone a template, dr start runs again in the cloned directory.

If you're already in a repository but no start command or quickstart script exists:

  ✓ Starting application quickstart process...
  ✓ Checking DataRobot CLI version...
  ✓ Checking template prerequisites...
  ✓ Checking repository setup...
  → Finding and executing start command...

No start command or quickstart script found.
This template may not yet fully support the DataRobot CLI.
Please check the template README for more information on how to get started.

The command then exits; the setup wizard is not launched in this case.

Non-interactive mode

Skip all prompts and execute immediately:

dr start --yes

or

dr start -y

This is useful for:

  • CI/CD pipelines
  • Automated deployments
  • Scripted workflows

Using the alias

dr quickstart

Behavior

CLI version check and update

If your template defines a minimum CLI version in .datarobot/cli/versions.yaml, dr start checks the installed version before continuing.

If your CLI version does not meet the minimum, dr start prompts you to run dr self update.

Example prompt:

DataRobot AI Application Quickstart

  ✓ Starting application quickstart process...
  ✓ Checking DataRobot CLI version...

DataRobot CLI (minimal: v0.2.0, installed: v0.1.0)
Do you want to update it now?

Press 'y' or ENTER to confirm, 'n' to cancel

State tracking

The dr start command automatically tracks when it runs successfully by updating a state file with:

  • Timestamp of when the command last started (ISO 8601 format)
  • CLI version used

This state information is stored in .datarobot/cli/state.yaml within the repository. State tracking is automatic and transparent. No manual intervention is required.

The state file helps other commands (like dr templates setup) know that you've already run dr start, allowing them to skip redundant setup steps.

When task start exists

  1. task start command is detected in the Taskfile
  2. Command executes immediately (no confirmation prompt)
  3. Command completes when the task finishes
  4. State file is updated with current timestamp and CLI version

When a quickstart script exists (but no task start)

  1. Script is detected in .datarobot/cli/bin/
  2. User is prompted for confirmation (unless --yes or -y is used)
  3. If user confirms (or --yes is specified), script executes with full terminal control
  4. Command completes when script finishes
  5. State file is updated with current timestamp and CLI version

If the user declines to execute the script, the command exits gracefully and still updates the state file.

When no start command or script exists (and you're in a repository)

  1. No task start command is found in the Taskfile
  2. No executable quickstart script is found in .datarobot/cli/bin/
  3. The command displays a message that no start command or quickstart script was found and suggests checking the template README
  4. The command exits; the state file is not updated (nothing was run)

Note: If you're not in a DataRobot repository, the repository check step (before "Finding and executing start command") triggers the template setup wizard; that case is separate from "no start command or script."

Prerequisites checked

The "Checking template prerequisites" step verifies that tools required by the template (e.g. Task, Git) are installed. If any are missing, the command exits with an error.

Required tools and minimum versions can be configured in the template by creating .datarobot/cli/versions.yaml:

---
dr:
  name: DataRobot CLI
  minimum-version: 0.2.0
  command: dr self version
  url: https://github.com/datarobot-oss/cli
uv:
  name: uv Python package manager
  minimum-version: 1.7.0
  command: uv self version
  url: https://docs.astral.sh/uv/getting-started/installation/

The repository check runs before start-command detection. If the current directory is not within a DataRobot repository (no .datarobot/ directory), the command launches the template setup wizard instead of continuing to look for a start command.

Error handling

Not in a DataRobot repository

If you're not in a DataRobot repository, the command automatically launches the template setup wizard:

$ dr start
# Automatically launches: dr templates setup

No manual intervention is needed - the command handles this gracefully.

Missing prerequisites

$ dr start
Error: required tool 'git' not found

# Solution: Install the missing tool

Script execution failure

If a quickstart script fails, the error is displayed and the command exits. Check the script's output for details.

When to use dr start

✅ Good use cases

  • First-time setup—initializing a newly cloned template or starting from scratch.
  • Quick restart—restarting development after a break.
  • Onboarding—helping new team members get started quickly.
  • CI/CD—automating application initialization in pipelines.
  • General entry point—universal command that works whether you have a template or not.

❌ When not to use

  • Making configuration changes—use dr dotenv to modify environment variables.
  • Running specific tasks—use dr run <task> for targeted task execution.

See also

Tips

Creating a custom quickstart script

  1. Create the directory structure:

    mkdir -p .datarobot/cli/bin

  2. Create your script:

    # Create the script
cat > .datarobot/cli/bin/quickstart.sh <<'EOF'
#!/bin/bash
echo "Starting my custom quickstart..."
dr run build
dr run dev
EOF

  3. Make it executable:

    chmod +x .datarobot/cli/bin/quickstart.sh

  4. Test it:

    dr start --yes

Best practices

  • Keep scripts simple—focus on essential initialization steps.
  • Provide clear output—use echo statements to show progress.
  • Handle errors gracefully—use set -e in bash scripts to exit on errors.
  • Check prerequisites—verify .env exists and required tools are installed.
  • Make it idempotent—script should be safe to run multiple times.
  • Document behavior—add comments explaining what the script does.