cube_dbt

cube_dbt is the dbt integration for Cube that helps define the data model of the semantic layer on top of dbt models.

Installation

pip install cube_dbt

Usage

manifest_url = 'https://bucket.s3.amazonaws.com/manifest.json'
dbt = (
  Dbt
  .from_url(manifest_url)
  .filter(
    paths=['marts/'],
    tags=['cube'],
    names=['my_table', 'my_table_2']
  )
)
print(dbt.models)

# For use in Jinja templates:
print(dbt.model('name').as_cube())
print(dbt.model('name').as_dimensions(skip=['id']))
print(dbt.model('name').column('name').as_dimension())

Development

Run tests:

pdm run test

Preprocessing the manifest.json file

In case of a massive manifest file, it can be preprocessed for optimal performance. The cube_dbt package only reads the nodes dictionary where resource_type is model. Here's a list of all keys used by the cube_dbt package:

- nodes
  - name
  - path
  - description
  - config
    - materialized
    - tags
  - resource_type
  - database
  - schema
  - alias
  - relation_name
  - columns
    - name
    - description
    - data_type
    - meta
    - tags