Welcome to Docky’s documentation!
Docky is a Command Line Interface (CLI) tool written in Python (using CLY?!) for running Docker and Docker Compose commands.
Usage
To use Docky in your project, in your project’s root, run:
curl -fsSL https://raw.githubusercontent.com/mateusoliveira43/docky/main/install_docky.py | python -
This will copy Docky (latest version) files to the scripts
folder of your project.
Then add the project’s configuration in scripts/docky_cli/config.py
file.
To install (or update) Docky to a specific version, run:
curl -fsSL https://raw.githubusercontent.com/mateusoliveira43/docky/main/install_docky.py | python - <version>
You can check Docky help by running:
./scripts/docky.py

You can check a Docky command help by running:
./scripts/docky.py <command> --help
where <command>
is one of the commands presented in the Commands section
of Docky help.

To run your project with Docker for development, run:
./scripts/docky.py run

The command will create .env
file in project’s root (if it does not already exists),
Build the service’s images (and the images it depends on, if they are not already built),
and then connect to the service’s Container shell.
You can both run the commands inside the Container’s shell

or run each command individually, by running:
./scripts/docky.py run <command>
where <command>
can be a command and it’s arguments.

You can add Docky to the path, and call it with docky
instead of
./scripts/docky.py
by adding Docky to pyproject.toml
:
[tool.poetry.scripts]
docky = "scripts.docky_cli.__main__:CLI"
and running:
pip install -e ./