dtach 0.31.0

A Python tool to maintain a modular package architecture.

DTach

[!IMPORTANT] DTach is a fork of the unmaintained tach project.

Tach is a Python tool to enforce dependencies and interfaces, written in Rust.

Tach is inspired by the modular monolith architecture.

Docs

https://github.com/user-attachments/assets/11eec4a1-f80a-4f13-9ff3-91a9760133b6

Tach can enforce:

• 📋 Imports only come from declared dependencies
• 🤝 Cross-module calls use the public interface
• ⛓️‍💥 No cycles in the dependency graph

Tach is:

• 🌎 Open source
• 🐍 Installable via pip
• 🔧 Able to be adopted incrementally
• ⚡ Implemented with no runtime impact
• ♾️ Interoperable with your existing systems

Getting Started

Installation

uv (recommended)

uv add --dev dtach

pip

pip install dtach

Setup

Tach will guide you through initial project setup.

Run:

 tach init

After an introductory message, you will see a file tree interface allowing you to interactively configure your project.

Use the arrow keys to navigate, and mark each module boundary with 'Enter'. You can mark all of your top-level Python packages, or just a few that you want to track.

FAQ: What is a module?

If your Python code lives below your project root, or if you are working in a monorepo with multiple Python packages, mark your Python source roots using the 's' key.

Enforcement

Tach comes with a cli command to enforce the boundaries that you just set up! From the root of your Python project, run:

tach check

You will see:

✅ All modules validated!

You can validate that Tach is working by:

1. Removing an item from the depends_on key in tach.toml, or marking it as deprecated
2. By adding an import between modules that didn't previously import from each other.

Give both a try and run tach check again. This will generate an error:

❌ tach/check.py[L8]: Cannot use 'tach.filesystem'. Module 'tach' cannot depend on 'tach.filesystem'.

Each error indicates an import which violates your dependencies. If your terminal supports hyperlinks, click on the file path to go directly to the error.

When an error is detected, tach check will exit with a non-zero code. It can be easily integrated with CI/CD, Pre-commit hooks, and VS Code, and more!

Extras

Visualize your dependency graph.

tach show [--web]

Tach will generate a graph of your dependencies. Here's what this looks like for Tach:

Note that this graph is generated remotely with the contents of your tach.toml when running tach show --web.

If you would like to use the GraphViz DOT format locally, simply running tach show will generate tach_module_graph.dot in your working directory.

You can view the dependencies and usages for a given path:

tach report my_package/
# OR
tach report my_module.py

e.g.:

> tach report python/tach/filesystem
[Dependencies of 'python/tach/filesystem']
python/tach/filesystem/install.py[L6]: Import 'tach.hooks.build_pre_commit_hook_content'
python/tach/filesystem/project.py[L5]: Import 'tach.constants.CONFIG_FILE_NAME'
...
-------------------------------
[Usages of 'python/tach/filesystem']
python/tach/cache/access.py[L8]: Import 'tach.filesystem.find_project_config_root'
python/tach/cache/setup.py[L7]: Import 'tach.filesystem.find_project_config_root'
...

Tach also supports:

• Public interfaces for modules
• Deprecating individual dependencies
• Layered architecture
• Incremental adoption
• Generating JSON dependency maps
• Monorepos and namespace packages
• Domain ownership
• Inline 'ignore' comments
• Pre-commit hooks

More info in the docs. Tach logs anonymized usage statistics which can be opted out of. If you have any feedback, we'd love to talk!

If you have any questions or run into any issues, let us know by submitting a Github Issue!

---

Contributors