dbt-diagrams

This tool allows you to

1. generate an Entity Relationship Diagram (ERD) based on meta config in your dbt model docs. Currently it relies on Mermaid to render the output. You can choose to render your output inside your dbt docs directly, or store it as a separate SVG image.
2. Enable general Mermaid rendering capabilities to your dbt docs. Every Markdown code block tagged with mermaid will be rendered.

[!TIP] Read the blog post for more details

[!TIP] Demo Jaffle Shop docs pages

Installation instructions

As dbt-diagrams is just a Python package. Install it using your favourite Python package manager (e.g. pip install dbt-diagrams). In case you want to render your ERD to a SVG image, you will have to install the dbt-diagrams[svg] extras package as well.

Usage (1): add Mermaid rendering to dbt docs

Simply run dbt-diagrams docs generate instead of dbt docs generate. Any Markdown code block tagged with mermaid will now be picked up and rendered as an image.

Usage (2): specify ERD in meta blocks and render in dbt docs

This will achieve the same functionality as (1), plus the following: let's say you have the following models defined in your dbt project

version: 2

models:
  - name: customers
    description: >
      This table has basic information about a customer, as well as some derived facts based on a customer's orders

      ```mermaid[erd="customer_erd"]```
    config:
      meta:
        erd:
          connections:
            - diagram: customer_erd
              target: orders
              source_cardinality: one
              target_cardinality: one_or_more
              label: creates
    columns:
      - name: customer_id
        description: This is a unique identifier for a customer
      - name: first_name
        description: Customer's first name. PII.
      - name: last_name
        description: Customer's last name. PII.

  - name: orders
    description: >
      This table has basic information about orders, as well as some derived facts based on payments

      ```mermaid[erd="customer_erd"]```
    columns:
      - name: order_id
      - name: customer_id
        description: Foreign key to the customers table
      - name: order_date
        description: Date (UTC) that the order was placed
      - name: status
        description: '{{ doc("orders_status") }}'

Using the meta section of a model, you can define ERD connections to other models. Based on these connections and other table attributes the ERD can be generated. The target attribute is another dbt model name. Accepted relation cardinalities are one, zero_or_one, zero_or_more or one_or_more. Use the label attribute to specify a human readable interpretation to a relation. The diagram is optional and allows you to add a name to your ERD. This is useful in case you want to define multiple ERDs and reference them in dbt docs directly.

Notice the mermaid[erd="cusomer_erd"] expressions in the customers and orders model descriptions. When running dbt-diagrams docs generate, this will be replaced by the ERD Mermaid definition so that your ERD can be rendered in any dbt docs page.

erDiagram
   customers ||--|{ orders : creates

   customers {
    STRING customer_id
    STRING first_name
    STRING last_name
   }

   orders {
    STRING order_id
    STRING customer_id
    DATE order_date
    STRING status
   }

Usage (3): specify ERD in meta blocks and render as SVG

Given the same setup as above, you can also render your output to SVG:

1. Make sure you installed the dbt-diagrams[svg] extras. This will install a headless browser in which Mermaid can run.
2. Run dbt-diagrams render-erds -dbt-target-dir target --format svg --output ./out. This will use the manifest and catalog files from ./target to render all defined ERDs as SVG. All detected diagrams will be stored as SVG files in the ./out folder.

ERD Definition schema

Every erd section inside a meta block of a model will be picked up. It should look like the following:

erd:
  connections:
    - diagram: <Optional. This connection will be added to a diagram of this name>
      target: <Required. Other model name>
      source_cardinality: <Required. One of {zero_or_one, one, zero_or_more, one_or_more}>
      target_cardinality: <Required. One of {zero_or_one, one, zero_or_more, one_or_more}>
      label: <Optional. Any string that describes the relation from this model to target model.>