tedivm/paracelsus

Visualize SQLAlchemy Databases using Mermaid or Dot Diagrams.

Paracelsus

Paracelsus generates Entity Relationship Diagrams by reading your SQLAlchemy models.

• Paracelsus
  • Features
  • Usage
    • Installation
    • Basic CLI Usage
    • Importing Models
    • Generate Mermaid Diagrams
    • Inject Mermaid Diagrams
    • Creating Images
    • pyproject.toml
  • Sponsorship

Features

• ERDs can be injected into documentation as Mermaid Diagrams.
• Paracelsus can be run in CICD to check that databases are up to date.
• ERDs can be created as files in either Dot or Mermaid format.
• DOT files can be used to generate SVG or PNG files, or edited in GraphViz or other editors.

Usage

Installation

The paracelsus package should be installed in the same environment as your code, as it will be reading your SQLAlchemy base class to generate the diagrams.

pip install paracelsus

Basic CLI Usage

Paracelsus is primarily a CLI application.

paracelsus --help

It has three commands:

• version outputs the version of the currently installed paracelsus cli.
• graph generates a graph and outputs it to stdout.
• inject inserts the graph into a markdown file.

Importing Models

SQLAlchemy models have to be imported before they are put into the model registry inside of the base class. This is similar to how Alembic needs models to be imported in order to generate migrations.

The --import-module flag can be used to import any python module, which presumably will include one or more SQLAlchemy models inside of it.

paracelsus graph example_app.models.base:Base \
  --import-module "example_app.models.users" \
  --import-module "example_app.models.posts" \
  --import-module "example_app.models.comments"

The :* modify can be used to specify that a wild card import should be used. Make sure to wrap the module name in quotes when using this to prevent shell expansion.

paracelsus graph example_app.models.base:Base --import-module "example_app.models:*"

This is equivalent to running this style of python import:

from example_app.models import *

Generate Mermaid Diagrams

paracelsus graph example_app.models.base:Base --import-module "example_app.models:*"

erDiagram
 users {
  CHAR(32) id PK
  DATETIME created
  VARCHAR(100) display_name "nullable"
 }

 posts {
  CHAR(32) id PK
  CHAR(32) author FK
  TEXT content "nullable"
  DATETIME created
  BOOLEAN live "nullable"
 }

 comments {
  CHAR(32) id PK
  CHAR(32) author FK
  CHAR(32) post FK "nullable"
  TEXT content "nullable"
  DATETIME created
  BOOLEAN live "nullable"
 }

 users ||--o{ posts : author
 posts ||--o{ comments : post
 users ||--o{ comments : author

When run through a Mermaid viewer, such as the ones installed in the markdown viewers of many version control systems, this will turn into a graphic.

erDiagram
 users {
  CHAR(32) id PK
  DATETIME created
  VARCHAR(100) display_name "nullable"
 }

 posts {
  CHAR(32) id PK
  CHAR(32) author FK
  TEXT content "nullable"
  DATETIME created
  BOOLEAN live "nullable"
 }

 comments {
  CHAR(32) id PK
  CHAR(32) author FK
  CHAR(32) post FK "nullable"
  TEXT content "nullable"
  DATETIME created
  BOOLEAN live "nullable"
 }

 users ||--o{ posts : author
 posts ||--o{ comments : post
 users ||--o{ comments : author

Inject Mermaid Diagrams

Mermaid Diagrams and Markdown work extremely well together, and it's common to place diagrams inside of project documentation. Paracelsus can be used to inject diagrams directly into markdown configuration. It does so by looking for specific tags and placing a code block inside of them, replacing any existing content between the tags.

## Schema
<!-- BEGIN_SQLALCHEMY_DOCS -->

<!-- END_SQLALCHEMY_DOCS -->

paracelsus inject db/README.md example_app.models.base:Base --import-module "example_app.models:*"

The --check flag can be used to see if the command would make any changes. If the file is already up to date then it will return a status code of 0, otherwise it will return 1 if changes are needed. This is useful in CI/CD or precommit hook to enforce that documentation is always current.

paracelsus inject db/README.md example_app.models.base:Base --import-module "example_app.models:*" --check

Creating Images

GraphViz has a command line tool named dot that can be used to turn dot graphs into images.

To create an SVG file:

paracelsus graph example_app.models.base:Base --import-module "example_app.models:*" --format dot | dot -Tsvg > output.svg

To create a PNG file:

paracelsus graph example_app.models.base:Base --import-module "example_app.models:*" --format dot | dot -Tpng > output.png

pyproject.toml

The settings for your project can be saved directly in the pyprojects.toml file of your project.

[tool.paracelsus]
base = "example.base:Base"
imports = [
  "example.models"
]

Sponsorship

This project is developed by Robert Hafner If you find this project useful please consider sponsoring me using Github!

paracelsus's Issues

error ModuleNotFoundError: No module named 'tomllib' (python 3.10)

while using python 3.10, pip installs paracelsus==0.5.0 . Running pipenv run paracelsus graph gives error

ModuleNotFoundError: No module named 'tomllib'

tomllib is supported from 3.11. Maybe 'requires-python' should be specified or some other toml parser used.

How to run the cli when I don't have a package/module set up?

I have a python setup where I've got some relatively simple scripts that interact with a fairly complex database that I've wrapped with SQLAlchemy. Here's the structure:

working dir with script1.py, script2.py
|--- db/
|    |--- __init__.py (empty)
|    |--- models.py

No matter what I've tried to pass to paracelsus on the command line, I can't get it to load my base class. Any ideas on what I can do to use this tool?

Current typing format blocks python < 3.10.x

There is some type hinting in pyproject.py that uses the | operand, which wasn't introduced until 3.10.

Traceback (most recent call last):
  File "/redacted/bin/paracelsus", line 5, in <module>
    from paracelsus.cli import app
  File "/redacted/lib/python3.8/site-packages/paracelsus/cli.py", line 11, in <module>
    from .pyproject import get_pyproject_settings
  File "/redacted/lib/python3.8/site-packages/paracelsus/pyproject.py", line 11, in <module>
    def get_pyproject_settings(dir: Path = Path(os.getcwd())) -> Dict[str, Any] | None:
TypeError: unsupported operand type(s) for |: '_GenericAlias' and 'NoneType'