Dagster DG Core

1.13.0 · active · verified Sat Apr 11

dagster-dg-core is a foundational library within the Dagster ecosystem, primarily powering the `dg` command-line interface (CLI). It enables local development, project scaffolding, deployment to environments like Dagster Cloud, and programmatic interaction with Dagster entities such as assets, runs, and jobs via CLI commands. As of version 1.13.0, it supports advanced features including AI-assisted development tools and enhanced asset management capabilities. Dagster maintains a frequent release cadence, often issuing weekly or bi-weekly patches and regular minor version updates across its libraries.

Warnings

gotcha All Dagster ecosystem packages (e.g., `dagster`, `dagster-webserver`, `dagster-dg-core`, and other `dagster-*` libraries) should always be installed at the exact same version number to ensure compatibility and avoid runtime errors. Mismatched versions are a very common source of issues.
Fix: Ensure `pip install` commands specify consistent versions across all `dagster-*` packages, e.g., `pip install dagster==1.x.y dagster-dg-core==1.x.y`.
breaking As of Dagster core v1.12.18 (and corresponding `dagster-dg-core` versions), the `psycopg2-binary` package is no longer a transitive dependency of `dagster-postgres`. If your project relies on `dagster-postgres` (or integrations that use it) and implicitly used `psycopg2-binary`, you must now explicitly add `pip install psycopg2-binary` to your project dependencies.
Fix: Explicitly install `psycopg2-binary` in your environment: `pip install psycopg2-binary`.
gotcha Older versions of `dagster-dg-core` (e.g., v1.12.3) might have an over-conservative dependency pin on `typer < 0.17`. This can lead to conflicts with other Python packages that require a newer `typer` version (e.g., `>= 0.20.1`).
Fix: If encountering `typer` version conflicts, ensure you are on the latest `dagster-dg-core` version, which may have relaxed the pin, or manage `typer` explicitly if feasible in your dependency resolution strategy.

Install

pip install dagster dagster-dg-core Install with pip

Imports

cli
```
from dagster_dg_core import cli
```
Direct Python imports from `dagster_dg_core` for defining user-facing Dagster assets or ops are generally uncommon. This library primarily underpins the `dg` CLI tool. Most interaction is through command-line commands, rather than programmatic Python imports for data pipeline development.

Quickstart

This quickstart demonstrates how to initialize a new Dagster project and launch the development server using the `dg` CLI, which is powered by `dagster-dg-core`. This allows you to explore the Dagster UI and begin defining assets.

pip install dagster dagster-dg-core
mkdir my-dagster-project
cd my-dagster-project
dg new --name my_first_project --preset minimal

# Navigate into the newly created project directory
cd my_first_project

# Start the Dagster development server and UI
dg dev

view raw JSON →