import-cruiser 0.2.6

Analyze, validate, and visualize Python import dependencies — like dependency-cruiser for Python.

import-cruiser

Analyze, validate, and visualize Python import dependencies.

import-cruiser is a CLI tool for Python projects inspired by dependency-cruiser. It parses Python import statements, builds a dependency graph, detects circular dependencies, validates the graph against configurable rules, and can export the results as JSON or DOT (Graphviz) format.

---

Features

• 🔍 Parse all Python imports in a project directory
• 🔄 Detect circular dependencies automatically
• ✅ Validate dependencies against user-defined JSON rules
• 📊 Export dependency graphs as JSON or DOT/Graphviz
• 🖥️ CLI with three subcommands: analyze, validate, export

---

Project self-graph (SVG)

Current dependency graph of src/import_cruiser (generated by import-cruiser itself):

---

Installation

With pip

pip install import-cruiser

From source (with Poetry)

git clone https://github.com/kevin91nl/import-cruiser.git
cd import-cruiser
poetry install

---

CLI Usage

Quick start: generate HTML/SVG for your repo

Use this when you want a graph fast, without setting up rules first.

# 1) Install tool
pip install import-cruiser

# 2) From your repository root, export an interactive HTML graph
import-cruiser export . --format html --output deps.html

# 3) Export SVG directly
import-cruiser export . --format svg --output deps.svg

# 4) (Optional) Export DOT too, for custom Graphviz rendering
import-cruiser export . --format dot --output deps.dot

Useful filters for larger repos:

# Focus on src/ and skip tests
import-cruiser export . --format html --include-path '^src/' --exclude-path '/tests/' --output deps-src.html

# Focus on a package/module prefix
import-cruiser export . --format svg --include '^myapp\.' --output deps-myapp.svg

Tip: HTML is best for exploration (hover/pan/zoom), SVG is best for embedding in docs.

CI usage (GitHub Actions)

name: dependency-graph

on:
  workflow_dispatch:
  pull_request:

jobs:
  graph:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-python@v5
        with:
          python-version: '3.11'
      - run: pip install import-cruiser
      - run: import-cruiser export . --format html --output deps.html
      - run: import-cruiser export . --format svg --output deps.svg
      - uses: actions/upload-artifact@v4
        with:
          name: dependency-graph
          path: |
            deps.html
            deps.svg

Command overview

import-cruiser [OPTIONS] COMMAND [ARGS]...

Commands:
  analyze   Analyze imports and output results.
  export    Export the dependency graph.
  validate  Validate dependencies against rules.

analyze

Scan a Python project and output a JSON or DOT dependency report.

# JSON report (default)
import-cruiser analyze ./myproject

# DOT format for Graphviz
import-cruiser analyze ./myproject --format dot

# Write to file
import-cruiser analyze ./myproject --output report.json

JSON output structure:

{
  "summary": {
    "modules": 12,
    "dependencies": 18,
    "cycles": 0,
    "violations": 0
  },
  "modules": [...],
  "dependencies": [...],
  "cycles": [],
  "violations": []
}

validate

Validate dependencies against rules defined in a JSON configuration file.

import-cruiser validate ./myproject --config import-cruiser.json

# Exit non-zero if there are any violations (useful in CI)
import-cruiser validate ./myproject --config import-cruiser.json --strict

# Emit linter-style output for editor/CI integrations
import-cruiser validate ./myproject --config import-cruiser.json --output-format flake8
import-cruiser validate ./myproject --config import-cruiser.json --output-format pylint
import-cruiser validate ./myproject --config import-cruiser.json --output-format github

Linter integration output modes

validate supports multiple output formats via --output-format:

• json (default): full graph + violations + cycles
• flake8: path:line:col: CODE message
• pylint: path:line: [CODE] message
• github: GitHub Actions annotations (::error file=...::...)

This makes it easy to plug import-cruiser into common linting pipelines. In practice, the most-used Python lint tools are typically Ruff, Flake8, and Pylint. For those stacks, flake8/pylint output modes are editor-friendly, and github is ideal for GitHub Actions logs.

export

Export the dependency graph to DOT format (compatible with Graphviz).

import-cruiser export ./myproject --output graph.dot

# Render with Graphviz
dot -Tsvg graph.dot -o graph.svg

---

Configuration

Create a import-cruiser.json file to define dependency rules:

{
  "rules": [
    {
      "name": "no-circular",
      "severity": "error",
      "from": { "path": "myapp\\.ui" },
      "to":   { "path": "myapp\\.data" },
      "allow": false
    },
    {
      "name": "allow-utils",
      "severity": "warn",
      "from": {},
      "to": { "path": "myapp\\.utils" },
      "allow": true
    }
  ],
  "options": {
    "include_external": false
  }
}

Rule schema

Field	Type	Required	Description
name	string	✅	Unique rule identifier
severity	string	✅	"error", "warn", or "info"
from	object	✅	Source module pattern (see Pattern object below)
to	object	✅	Target module pattern (see Pattern object below)
allow	boolean	❌	true (default) = allowed; false = forbidden

Pattern object

Field	Type	Description
path	string	Regular expression matched against the module name

An empty pattern object {} matches all modules.

---

Examples

Detect circular dependencies

import-cruiser analyze ./myproject
# Check the "cycles" key in the JSON output

Enforce layered architecture

{
  "rules": [
    {
      "name": "no-data-to-ui",
      "severity": "error",
      "from": { "path": "\\.data" },
      "to":   { "path": "\\.ui" },
      "allow": false
    }
  ]
}

import-cruiser validate ./myproject --config import-cruiser.json --strict

Visualize dependencies

import-cruiser export ./myproject --output deps.dot
dot -Tpng deps.dot -o deps.png
open deps.png

---

Development

# Install dev dependencies
poetry install

# Run tests
poetry run pytest

# Run tests with coverage
poetry run pytest --cov

Architecture enforcement (pre-commit)

This repository uses import-cruiser itself in pre-commit to enforce dependency boundaries:

• Config file: import-cruiser.json
• Hook: import-cruiser-architecture
• Command: PYTHONPATH=src python3 -m import_cruiser.cli validate src --config import-cruiser.json --strict --output-format pylint

This gives linter-friendly output and fails commits when architectural rules are violated.

---

License

MIT