Getting Started with the Daco CLI

The Daco CLI is the fastest way to start building data products with OpenDPI. This guide covers installation, project initialization, connection setup, port definition, and schema translation.

Installation

macOS / Linux (Homebrew)

brew install dacolabs/tap/daco

Windows (Scoop)

scoop bucket add dacolabs https://github.com/dacolabs/scoop-bucket.git
scoop install daco

Go

go install github.com/dacolabs/cli/cmd/daco@latest

Verify the installation:

daco version

Initialize Your First Data Product

Create a new directory and initialize a project:

mkdir my-data-product
cd my-data-product
daco init

This generates an opendpi.yaml file with basic metadata:

opendpi: "1.0.0"

info:
  title: My Data Product
  version: "0.1.0"
  description: ""

Add a Connection

Connections tell Daco where your data lives. Run:

daco connections add

The CLI walks you through selecting a connection type (PostgreSQL, Kafka, S3, BigQuery) and configuring the required variables. Your opendpi.yaml gains a connections section:

connections:
  analytics_db:
    type: postgresql
    host: analytics.db.example.com
    variables:
      database: analytics
      schema: public

Define a Port

Ports are the datasets your data product exposes. Add one:

daco ports add

This creates a schema file in a schemas/ directory and registers the port in your opendpi.yaml:

ports:
  events:
    description: "Raw user events"
    connection: analytics_db
    location: public.events
    schema:
      type: object
      required: [event_id, user_id, timestamp]
      properties:
        event_id:   { type: string }
        user_id:    { type: string }
        event_type: { type: string }
        timestamp:  { type: string, format: date-time }

Translate Your Schema

This is where it gets useful. Translate your port schema to any supported format:

# PySpark
daco ports translate --format pyspark events

# dbt model YAML
daco ports translate --format dbt events

# Pydantic model
daco ports translate --format pydantic events

# Go struct
daco ports translate --format go events

Supported formats include Avro, Protobuf, Pydantic, Python, Go, Scala, Spark SQL, Spark Scala, PySpark, Databricks SQL, Databricks Scala, Databricks PySpark, DQX YAML, and Markdown — over 12 in total.

What's Next

• Add more ports to model your full data product interface
• Run daco validate to check your schema against the OpenDPI specification
• Push to your repository and connect Daco Studio to get a live catalog